University of Southern California Dissertations and Theses

Germanium-silicon fractionation in a continental shelf environment: insights from the northern Gulf of Mexico

(USC Thesis Other)

Germanium-silicon fractionation in a continental shelf environment: insights from the northern Gulf of Mexico

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
GERMANIUM-SILICON FRACTIONATION IN A
CONTINENTAL SHELF ENVIRONMENT:
INSIGHTS FROM THE NORTHERN GULF OF MEXICO

by
Jotautas Jokūbas Baronas

December 2014

A Thesis Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(GEOLOGICAL SCIENCES)

1

Acknowledgements

I would like to express my utmost gratitude to Dr. Douglas Hammond for his trust, guidance, and
patience throughout my learning process that has resulted in this thesis. I would also like to thank Dr.
Josh West, as well as the rest of the Hammond-West group, for always taking the time to engage in
insightful discussions or to help out in the lab. Will Berelson, Silke Severmann, and especially James
McManus are thanked for their support during the cruise, selflessly sharing equipment, samples, data, and
ideas. The crew and the science team of R/V Endeavor made it possible to collect the samples. Nick
Rollins and Christa Wolfe have helped with sample collection and analysis. Jim Moffett and Sergio
Sañudo-Wilhelmy are thanked for providing me with access to their ICP-MS instrument.
This work was mainly supported by the U.S. National Science Foundation (NSF) grant OCE-1061700
awarded to Doug Hammond. Ship time for the cruise was supplied by the NSF grant OCE-1029878
awarded to Will Berelson. Conference travel funding was generously provided by the Earth Sciences
Department (Dornsife College of Letters, Arts and Sciences, University of Southern California) and the
USC Graduate Student Government.
Finally, I would like to express my sincere thankfulness and love to my family (ačiū Mama ir Gusti!)
and my friends, whether here or far away - you know who you are.

2

Table of Contents

Acknowledgements 1
List of Tables 3
List of Figures 4
Abstract 5
1. Introduction 7
2. Study site 11
3. Methods 11
3.1 Water and sediment sampling 11
3.2 Bulk sediment leaching 13
3.3 Biogenic silica cleaning and dissolution 14
3.4 Nutrient and metal analysis 15
3.5 Germanium analysis 15
4. Results 16
4.1 Water column 16
4.2 Pore waters 17
4.3 Benthic fluxes 19
4.4 Sediments 20
5. Discussion 21
5.1 Water column and biogenic fractionation 21
5.2 Ge/Si fractionation based on mass balance for shelf water and sediments 24
5.3 Sediment diagenesis 30
5.4 Spatial extent and variation of the non-opal Ge sink 40
6. Conclusions 41
7. Tables and Figures 43
8. References 59
Appendix A: Additional data and figures 68
Appendix B: ID-HG-ICP-MS Ge data workup 75

3

List of Tables

Table 1. Sampling station details .......................................................................................................... 43
Table 2. Benthic Si and Ge fluxes ........................................................................................................ 44
Table 3. Summary of sediment results and characteristics ................................................................... 45
Table 4. Water column data from Sta. 1 used in the box model........................................................... 46
Table 5. Box model input parameters and ranges tested ...................................................................... 47
Table 6. Box model results ................................................................................................................... 48
Table 7. Comparison of Si and Ge fluxes at different locations ........................................................... 49

4

List of Figures

Figure 1. Map of sampling stations of cruise EN-497 that are discussed in this study. .............................. 50
Figure 2. Water column profiles of Ge and Si concentrations, as well as Ge/Si ratios, over the continental
shelf (Sta. 1 and 2), continental slope (Sta. 5), and offshore (Sta. G) ......................................................... 51
Figure 3. Ge and Si relationship in the water column of the Northern Gulf of Mexico ............................. 52
Figure 4. Pore water profiles of dissolved Si, Ge, Fe
2+
, Mn
2+
, and NH
4
+
at Sta. 1, Sta. 2, Sta. M, and Sta. G
.................................................................................................................................................................... 53
Figure 5. Pore water profiles of dissolved Si, Fe
2+
, Mn
2+
, and NH
4
+
at several stations. ............................ 54
Figure 6. Core incubation experiment data ................................................................................................. 55
Figure 7. Si depletion and enrichment over the shelf.................................................................................. 56
Figure 8. Box model of Sta. 1 water column and sediment Si and Ge cycling ........................................... 57
Figure 9. Pore water Ge/Si plotted as a function of 1/Si ............................................................................. 58

5

ABSTRACT
Germanium (Ge) is a trace element whose biogeochemical behavior closely resembled that of silicon
(Si). Both elements are supplied to the ocean primarily by riverine and hydrothermal inputs. Once in the
ocean, they get consumed by silicifying organisms and exported to the sediments as biogenic silica (bSi).
The Ge/Si ratio recorded in sedimentary bSi covaries with glacial-interglacial cycles as far back as the
Mid-Miocene (Shemesh, A., Mortlock, R.A., and Froelich, P.N., 1989, Late cenozoic Ge/Si record of
marine biogenic opal: Implications for variations of riverine fluxes to the ocean. Paleoceanography, 4 (3),
221–234), suggesting important changes in Ge and likely Si cycling over millennial timescales. However,
the intensity of terrestrial rock weathering, variations in the size of the marine non-opal Ge sink, and
biological fractionation by marine silicifiers could all potentially be used to explain the Ge/Si paleorecord.
Here we present a study investigating the effects of biological fractionation and sediment diagenesis on
Ge cycling in the northern Gulf of Mexico – a continental margin strongly influenced by the Mississippi-
Atchafalaya river plume and rapidly accumulating sediments. We have measured Ge and Si
concentrations in the water column, sediment pore waters, and biogenic silica at several stations on the
continental shelf, as well as deeper waters. Results indicate extensive biological fractionation by silicifiers
in the water column, elevating the shelf surface water Ge/Si up to 2-25 µmol/mol relative to the global
ocean (0.76 µmol/mol). A simple steady state box model was built to constrain the Ge and Si budget in
the water column and the sediments of one station. Biological fractionation parameters were estimated
using a Michaelis-Menten kinetic model and agree well with previous observations from the global ocean.
Based on the box model results, as well as pore water and sediment analyses, the GOM sediments act as a
weak but significant non-opal Ge sink, comprising 2-49% of total Ge burial flux. There is evidence for
both authigenic aluminosilicate formation (i.e., reverse weathering) and iron sulfide formation in these
sediments. Therefore, either of these processes could be responsible for Ge non-opal sequestration in this
environment. A summary of calculated opal and non-opal Ge fluxes from previously published data
shows high spatial and temporal variability of the non-opal Ge sink. The non-opal Ge flux into GOM
sediments is around an order of magnitude lower than that into California Borderlands basins sediments.

6

Overall, our results 1) provide further evidence for biological Ge/Si fractionation by diatoms and 2)
suggest that the non-opal Ge sink may be not as efficient as previously thought but instead extend over a
larger portion of the seafloor, including deep ocean sediments. Further investigation of the Ge cycle is
needed to explain the glacial-interglacial Ge/Si paleorecord.

7

1. INTRODUCTION
Germanium (Ge) is a trace element whose inorganic species’ chemical behavior in the natural
environment closely resembles that of silicon (Si). Both elements are supplied to the ocean as dissolved
Si(OH)
4
and Ge(OH)
4
species in river water, the result of continental silicate rock weathering, as well as
in hydrothermal fluid discharge (Froelich et al., 1985a).Together these fluxes account for ~70% of total
Ge input, as calculated from the most recent global Si budget (Tréguer and De La Rocha, 2013) and the
best Ge/Si estimates for global river (Froelich et al., 1992) and hydrothermal (Mortlock et al., 1993;
Wheat and McManus, 2005; Wheat and McManus, 2008) discharge into the ocean. Dust delivery,
seafloor basalt weathering, terrestrial biogenic silica, and groundwater inputs are believed to account for
the rest but their magnitude is less well constrained. Ge and Si are removed from the ocean through
uptake by silicifiers (mainly diatoms in surface seawater and sponges in the benthos). Most of the
biogenic silica (alternatively, bSi or opal) formed this way re-dissolves as it sinks into the deep ocean, but
some is buried in the sediments. While different sources have distinct Ge/Si ratios, and these ratios also
differ in regional sinks, the oceanic residence time of each element is long compared to ocean mixing
rates (Hammond et al., 2000). This, coupled with relatively similar dynamics in biological cycling has
resulted in very similar depth profiles of each element (Froelich et al., 1985a) and a constant dissolved
Ge/Si
SW
ratio of 0.76 µmol/mol in the deep ocean (Sutton et al., 2010).
Ge/Si
opal
recorded in biogenic silica preserved in marine sediments has varied significantly over time
(Kolodny and Halicz, 1988; Murnane and Stallard, 1988; Froelich et al., 1989; Shemesh et al., 1989;
Mortlock et al., 1991; Froelich et al., 1992; Bareille et al., 1998; Lin and Chen, 2002). Most notably
Ge/Si
diatom
shows both a long-term Cenozoic trend (Shemesh et al., 1989) and rapid changes
corresponding to glacial-interglacial cycles over the last 500 ky (Mortlock et al., 1991). Initially, the
glacial-interglacial Ge/Si variations were interpreted as a biological fractionation signal (similar to the
more recently developed δ
30
Si proxy (see the review of Hendry and Brzezinski (2014)), thus providing
information on paleo-Si concentrations and diatom ecology in the past (Froelich et al., 1989). While it is

8

well established that sponges strongly discriminate against Ge (Ellwood et al., 2006), there has been a lot
of contention whether that is the case with diatoms, with evidence both in support of (Shemesh et al.,
1989; Froelich et al., 1989; Murnane and Stallard, 1988; Ellwood and Maher, 2003) and against (Froelich
et al., 1992; Bareille et al., 1998) biological fractionation. Recent field measurements have shown that
diatom Ge/Si deviates from that of the surrounding seawater only at Si concentrations below 10 µM
(Sutton et al., 2010). This is also suggested by preliminary lab culture experiments (Sutton, 2011). The
contradictory evidence previously observed is most likely a result of the interplay of diffusion-mediated
and active transporter-mediated Si uptake mechanisms employed by diatoms (Thamatrakoln and
Hildebrand, 2008). Nevertheless, as argued by Sutton et al. (2010), paleorecords of diatom Ge/Si from the
Southern Ocean presented by Mortlock et al. (1991) should reflect whole ocean Ge/Si changes as opposed
to local changes in biological fractionation, since the Si concentration of surface water above these
Southern Ocean coring sites is unlikely to ever have been lower than 20 µM.
Whole ocean glacial-interglacial Ge/Si variations imply a rapid (<10 ky, the approximate residence
time of Ge in the ocean (Hammond et al., 2000)) change in either the input or the output fluxes of either
Ge or Si. Considering the two dominant input sources, dissolved Si and Ge derived from continental
weathering (comprising ~50% of total Si and ~30% of total Ge input; calculated from the recently revised
Si budget (Tréguer and De La Rocha, 2013) and the global mean riverine Ge/Si (Froelich et al., 1992)) is
most likely to have changed due extreme changes in climate. There is no evidence suggesting that
hydrothermal flux, which accounts for ~50% of Ge and ~10% of Si input, may have varied over such
geologically short timescales. Other inputs of both elements, such as dust deposition and groundwater
discharge, are minor and unlikely to influence the whole ocean signature.
With this in mind, several efforts have been made to interpret Ge/Si as a proxy for chemical
weathering intensity on the continents (Murnane and Stallard, 1990; Froelich et al., 1992; Lin and Chen,
2002). There have been a number of studies investigating Ge/Si systematics in rivers and soils (Mortlock
and Froelich, 1987; Murnane and Stallard, 1990; Froelich et al., 1992; Chillrud et al., 1994; Kurtz et al.,

9

2002; Anders et al., 2003; Derry et al., 2006; Lugolobi et al., 2010) and as a result the behavior of Ge in
the weathering environment is relatively well understood. In the global dataset, there is a robust negative
correlation of river Ge/Si with Si concentration (Froelich et al., 1992), indicating that silicate weathering
intensity exhibits the first order control on Ge/Si. In certain environments, the biological Si cycling can
“overprint” the weathering signal since terrestrial plants tend to strongly discriminate against Ge during
growth (Blecker et al., 2007; Delvigne et al., 2009). This fact has been exploited by a number of studies to
trace the nature of soil and river Si sources (Derry et al., 2005; Cornelis et al., 2010; Opfergelt et al.,
2010; Sparks et al., 2010; Lugolobi et al., 2010; White et al., 2012; Cornelis et al., 2014) but it is unclear
how significant this effect is on a global scale.
However, a straightforward interpretation of the diatom Ge/Si paleorecord as reflecting a terrestrial
signal has been rendered impossible by the discovery of the non-opal Ge sink. Decoupling of Ge and Si
chemical behavior during marine sediment early diagenesis has been observed in multiple studies
(Murnane et al., 1989; Hammond et al., 2000; King et al., 2000; McManus et al., 2003), where Ge
released from bSi dissolution gets scavenged during authigenic mineral formation in Fe-rich anoxic
sediments. Iron (oxy)hydroxides (FeOx), pyrite, and Fe-rich aluminosilicates have all been suggested as
potential candidates for this role (Murnane et al., 1989; Hammond et al., 2000; King et al., 2000).
Hammond et al. (2004b) have shown that a drop in seawater temperature, which would slow down the
dissolution rate of sinking bSi, and in turn increase the amount of Ge delivered to the sediments to be
diagenetically sequestered, could explain the whole ocean Ge/Si difference observed between the
Holocene and the Last Glacial Maximum. However, the exact diagenetic processes responsible have not
been identified, and the spatial extent of this non-opal Ge sink in the modern environment is poorly
constrained.
Continental margin environments often experience high rates of biogenic and lithogenic particle
sedimentation, thus accounting for 50-80% of global C and Si burial, while comprising only 15-20% of
the seafloor area (Aller, 2014). These same conditions (high delivery rate of both biogenic and terrestrial

10

particles) are conducive for the formation of reducing sediments and potential diagenetic Ge
sequestration.
Additional motivation for studying river plume environments comes from the fact that they serve as
the entry zone and a “filter” for the riverine Si and Ge entering the ocean. Most of this Si and Ge gets
taken up by diatom blooms in the river plume, exported to the coastal sediments, redissolved in the top
few centimeters of sediment, and refluxed back into the water column; possibly going through several
such cycles before being exported into the deep ocean. Therefore, any diagenetic processes in the
sediments can have a big effect on the final flux of these elements into the open ocean. For example, the
most recent review of the Si cycle (Tréguer and De La Rocha, 2013) estimates that 11 – 38% of river-
delivered Si might be buried during diagenetic secondary clay formation (i.e. “reverse weathering”) in
coastal margin sediments, indicating that such environments might play a key role in the Si cycle, despite
comprising only a small fraction of the global ocean floor.
With this in mind, we have investigated the Ge and Si cycling systematics in the water column and
the sediments of the northern Gulf of Mexico (GOM) – a continental margin setting heavily influenced by
the Mississippi-Atchafalaya river plume. The input of dissolved Si and dissolved and particulate Fe
results in intense Si and Fe cycling in near-river sediments, therefore this is a region with a high potential
for diagenetic (non-opal) sequestration of Ge. We have measured Ge/Si ratios in the water column,
sedimentary pore waters and biogenic silica, as well as the benthic flux of Ge and Si at multiple sites from
the continental shelf to the abyssal plane. Based on the data and box modelling, we have estimated the
rate of non-opal Ge burial in the northern Gulf of Mexico sediments and compared this to other
environments (California margin and the Southern Ocean), where this phenomenon has been previously
described.

11

2. STUDY SITE
The northern Gulf of Mexico is characterized by the large shallow Louisiana-Texas continental shelf,
onto which the Mississippi-Atchafalaya river system delivers a large amount of terrigenous sediment, as
well as high nutrient concentrations (Turner et al., 2007), causing extensive diatom blooms and bottom
water hypoxia over the shelf, especially during the summer months (Rabalais et al., 2002). Due to
prevailing winds, the plume most often gets trapped on the shelf and transported westwards (Walker et
al., 2005). Indeed, surface water salinity data indicate some plume influence as far as Station 2 (Fig. 1).
As mapped by Rabalais and Turner (2011) the week before we acquired our samples, the hypoxic bottom
water zone extended across most of the Louisiana shelf, encompassing the majority of the shelf, while
deeper water stations have not been affected (see Fig. 1 and Table 1).
Sediments in our study area are primarily composed of sand or silt-dominated clastic material
(Holmes, 1976) with organic carbon and calcite comprising around 0.5-2% and 2-20% of total mass
within the top 30 cm, respectively (Gordon and Goñi, 2004). bSi constitutes 0.5-1% of sediment mass in
the vicinity of the Mississippi river plume (Rabalais et al., 1996; Presti and Michalopoulos, 2008), which
we found to be true for the abyssal sediments as well.
3. METHODS
3.1 Water and sediment sampling
All samples were collected on research cruise EN-497 aboard R/V Endeavor during July 30, 2011 –
August 19, 2011, except for Mississippi River water which was collected using a plastic bucket a few
days later off a pier in New Orleans (Table 1).
Water column samples were collected using either a CTD rosette equipped with either Niskin or Go-
Flo bottles, and filtered through 0.45 µm filters immediately.

12

Sediment cores for pore water extraction and core incubations were collected using a multicorer and
immediately placed in a cold room at near in-situ temperature. Pore waters were extracted either by
slicing the core into 3-5cm intervals under an N
2
atmosphere, centrifuging in 50 mL polypropylene tubes,
and filtering through 0.45 µm filters (Sta. 1) or using Rhizons (Sta. M and Sta. G) within 1-3 days after
core retrieval. Aliquots for Ge analysis were acidified with trace metal clean-HCl.
Core incubations were done as previously described by (Hammond et al., 2004a). Briefly, sediment
cores with approx. 0.5 – 1 L of overlying water and a surface area of 70 cm
2
were capped and the water
stirred (16-40 RPM) at near in-situ temperature for several days, using a motor coupled to a magnetic
stirbar suspended below the cap. Aliquots of 30 mL were sampled through a plastic syringe several times
over the course of incubation, as a piston was advanced to keep air out. Incubations could not be done
exactly at in situ temperatures, because the wide range in bottom water temperature, coupled with the
need for long incubations (about 4 days) did not allow the cold van temperature to be readjusted to
precisely match all of our stations. A correction factor of 6.0%°C
–1
was applied to measured Si and Ge
fluxes (see Hammond et al. (2004a) for details).
Benthic chamber incubations were carried out as described previously (Berelson and Hammond,
1986). Each free-vehicle lander had 3 cylindrical polyvinyl chloride (PVC) chambers (30 cm i.d.,
identified as blue, red, or yellow) that were embedded in the sediment an hour or so after the device
landed on the sea floor. A hinged lid (acrylic) then closed, trapping a column of water (about 10 cm high)
in contact with a sediment area of 730 cm
2
. A suspended paddle stirred the overlying water at 7 RPM.
During the deployment, a known volume of trace metal–clean CsBr was injected into each chamber; the
dilution of this spike after 1 h was used to determine the volume of trapped water. Five or six water
samples were drawn during the deployment and stored in polyethylene containers until the lander was
recovered. At the end of each deployment (between 6 and 50 hours), a surface buoy attached to the lander
was retrieved and used to pull up the lander. Event timing for sampling and other functions was
programmed into a computer (Tattle- tale model 2B) housed in a pressure case on the lander. After

13

recovery, water samples were filtered (0.4 µm) and stored in a refrigerator until analysis for oxygen,
nutrients, and trace metals.
3.2 Bulk sediment leaching
The biogenic silica fraction in the sediments (%bSi, defined as SiO
2
) and Ge/Si
opal
was determined
using a wet chemical extraction technique adapted from (DeMaster, 1981; Mortlock and Froelich, 1989).
About 100 mg of oven-dried (60 °C) bulk sediment was ground using a mortar and pestle and placed in a
centrifuge tube. Some samples were then pre-treated with 5 mL of 10% H
2
O
2
and 5 mL of ~1M HCl to
remove organic and carbonaceous material, while others were not (Table 3) and then leached in 45 mL
5% Na
2
CO
3
. The tubes were shaken and centrifuged before drawing an aliquot, and again shaken after.
Two slightly different approaches were tested. Approach #1: Samples were leached at 85° C and 180 µL
aliquots taken at 0, 3, 4, and 5 h in duplicate and diluted with 2.5 mL 0.09 M HCl to obtain a final slightly
acidic solution. For Ge analysis, 5 mL aliquots of the 5 h sample were taken and diluted with 15 mL 0.4
M HCl. Si concentration in the aliquots was measured using the molybdate-blue method (standards were
matrix-matched). Clay dissolution was corrected for by fitting a linear regression to the 3-5 h increase of
Si concentration in the leachate solution and extrapolating the slope to the Si concentration intercept at
time=0, which corresponds to weight percent of bSi in sediments (%bSi; reported for mol. weight = 60
g/mol, water content disregarded). Approach #2: Samples were leached at 60° C, and aliquots were taken
at 0, 0.5, 1, 1.5, 2, 3, 4, 5, 6, 7, 12, and 24 h. Si was measured at each timepoint, while Ge was measured
at 3, 5, and 24h. The intercept of 12-24h Si concentration slope was used to correct for clay dissolution
and to determine %bSi.
Ge/Si
opal
was determined by measuring Ge concentration (see below) in the 5 h (Approach #1) or 24 h
(Approach #2) leachate solution and correcting for clay contribution using the 3-5 h or 12-24 h Si slope,
respectively, and assuming Ge/Si
lithogenic
= 2 µmol/mol (Mortlock and Froelich, 1987).

14

A blank control and our internal reference standard (sediment trap material from San Pedro basin)
were subjected to the same procedures. The control showed no leaching of either Si or Ge from the tube
walls, although the Na
2
CO
3
solution had an inherent blank that needed to be corrected for. The %bSi
determined for the internal reference standard was within 10% of previously measured value (Collins et
al., 2011) and agreed within 1% between Approach #1 and Approach #2. Using Approach #2 we
measured Ge/Si
opal
= 0.71 µmol/mol, which is in excellent agreement with Ge/Si
SW
in the San Pedro
basin, as well as the global ocean Ge/Si
SW
.
3.3 Biogenic silica cleaning and dissolution
To extract bSi and directly measure its Ge/Si ratio, it was isolated from bulk sediment and cleaned
using a procedure adapted from previous studies (Shemesh et al., 1988; Morley et al., 2004; Andersen et
al., 2011). Briefly, 60-100 g of wet sediment was heated to 90 °C for 5 h with 100 mL of 1 N HCl and
10% H
2
O
2
each, and left to settle overnight. Clays were disaggregated by adding 200 mL of 5% sodium
hexametaphosphate, heating to a boil and sonicating for a few minutes. The samples were then sieved
through a 20 µm sieve and thoroughly washed with DIW. Samples retained on the sieve were visually
inspected under an optical microscope and where a significant amount of non-bSi particles (e.g.
foraminifera or clays) was observed, the procedure was repeated using a smaller amount of reagents. The
samples were then dried in an oven and heavy-liquid separated using sodium polytungstate (~2.15 g/cm
3

density) to remove clastic particles. The floating fraction (bSi) was collected, thoroughly washed with
DIW and sieved again through a 20 µm sieve, then boiled in 30 mL of 25% H
2
O
2
for 3 h to remove any
remaining organic coating, thoroughly rinsed with DIW, and dried in an oven at 70° C. This procedure
yielded 10-40 mg of pure white bSi. SEM imaging showed that this cleaned fraction contained very few
clastic particles.
Approximately 1 mg of purified bSi was dissolved in 50 mL of 1% trace-clean Na
2
CO
3
solution,
heated to 90° C for several hours, and sampled while warm. This solution was filtered (0.45 µm) to

15

eliminate any particles. The Si concentration was calculated from the initial mass of bSi used. Ge
concentrations were determined using methods described below.
3.4 Nutrient and metal analysis
Silicic acid and NH
4
+
concentrations were measured using standard colorimetric techniques
(molybdate-blue (Mullin and Riley, 1955) for Si and salicylate-hypochlorite (Bower and Holm-Hansen,
1980) for NH
4
+
), detailed description by Hammond et al. (1996). Pore water and core incubation samples
were analyzed onboard during the cruise, while silicic acid concentration in the water column samples
was measured back in the lab. Fe, Mn, and other trace and major element concentrations were measured
using ICP-MS and ICP-AES at Oregon State University.
3.5 Germanium analysis
Inorganic Ge concentration was measured using isotope dilution-hydride generation-inductively
coupled plasma-mass spectrometry (ID-HG-ICP-MS) as described in detail by Mortlock and Froelich
(1996) and Hammond et al. (2000), on a Thermo Element 2 mass spectrometer at the University of
Southern California. Briefly, up to 0.5-15 mL sample is spiked with a known amount of
70
Ge solution to
obtain a
70
Ge/
74
Ge ratio of 2-20 (aiming for 4). TRIS (0.75 mL of 1.5 M, pH 6) and EDTA (0.25 mL 0.2
M) solutions are added and the mixture is diluted to 20 mL with DIW. Samples are equilibrated overnight
or longer. Sodium borohydride (1.5 mL 1M stabilized in 40 mM NaOH solution) is used to convert the
germanic acid to germane (GeH
4
), which is then transported using He as a carrier gas and collected in a
glass trap packed with silanized glass wool and immersed in liquid N
2
. Once the trap is warmed, GeH
4

volatilizes and is carried into the ICP-MS. The organic monomethyl- and dimethyl-Ge hydride species are
separated due to having higher boiling temperatures, producing separate chromatographic peaks. We did
not make an attempt to measure the concentration of these species. For inorganic Ge, we managed to
improve the sensitivity ~10× relative to that achieved previously (Hammond et al., 2000) by introducing
an additional Ar carrier gas flow prior to the sample entering the torch.

16

Ge peaks at m/z = 70 and m/z = 74 are integrated to obtain the total signal for the respective Ge
isotopes. Ge concentration is then calculated using the known amount of sample and
70
Ge spike added.
Several blanks are spiked and measured using the same procedure along with each batch of samples. Most
of the Ge blank comes from the reagents and is in the 10-50 fmol range (a typical sample contains ~300
fmol of natural Ge). Instrumental mass discrimination is determined separately each day by running
unspiked standards, and ranges from 0 to 3.5% / amu. The data presented here was obtained on 7 days of
analysis throughout the course of 2 years. The reproducibility of
70
Ge spike concentration was checked
using internal standards and is better than 3% (1σ S.D., n = 39). The reproducibility of several internal
seawater standards, as well as a NIST 3120a Ge standard is in the 1-3 % range (1σ, n = 34 total). The
reproducibility of a surface seawater sample with Ge concentration < 2 pM was 0.5 pM or 17% (1σ, n =
7). We thus adopt the conservative uncertainty of 3% or 0.5 pM, whichever is higher.
4. RESULTS
4.1 Water column
The Mississippi river Si and Ge concentrations were 169 µM and 264 pM, respectively, resulting in
Ge/Si of 1.56 ± 0.07 µmol/mol. Shelf and offshore water column concentrations of these elements were
much lower – the profiles of several stations are given in Fig. 2 (all values listed in Table A1 of the
Appendix). Profiles generally show depletion in surface water and enrichment in bottom water for both
elements, indicating diatom productivity in the photic zone and biogenic silica remineralization below.
However, there are significant differences between Ge and Si concentration variations with depth,
especially in the Si-depleted surface waters (concentrations below 5 µM; Fig. 3a).
GOM shelf water column Ge/Si
SW
ratio varies with depth at each station and among stations, with
values ranging from 0.60 ± 0.03 to 25.1 ± 5.5 µmol/mol (excluding a single uncharacteristically low value
of 0.23 ± 0.06 µmol/mol; Fig. 2 and 3a, Table A1 in Appendix A). The highest ratios are observed in the
most Si-depleted surface waters at the shallowest stations (Sta. 1 and 2). Individual deep water (below

17

500 m) samples at Station G exhibit relatively constant Ge/Si
SW
ratios in the range of 0.60 ± 0.03 to 0.86
± 0.04, with an average of 0.79 ± 0.11 µmol/mol. A linear regression of Ge vs Si concentrations at this
station produces a slope of Ge/Si
SW
= 0.70 ± 0.06 µmol/mol (Fig. 3b). Both of these values are
indistinguishable from the global ocean regression value of 0.76 µmol/mol (Sutton et al., 2010). In
contrast, all stations on the shelf show Ge/Si
SW
ratios that are significantly higher than the global ocean,
throughout the entire water column, with Sta. 2 exhibiting some of the highest Ge/Si
SW
values ever
measured (Figs. 2 and 3). Linear regression of all the shelf station data results in a Ge/Si slope of 2.16 ±
0.08 µmol/mol – much higher than global ocean value – albeit with significant scatter (Fig. 3b).
4.2 Pore waters
Based on their Si and NH
4
+
pore water profiles, all the stations can be divided into four categories:
Type A. Continental shelf (15-30 m water depth) in the vicinity of either the Mississippi or Atchafalaya
river plumes (Sta. 1, 4, and A);
Type B. Continental shelf closer to the coast and generally further away from river plumes (Sta. 2, 3, 9,
and 10);
Type C. Continental slope (50-500 m water depth; Sta. 5, 7, and M); and
Type D. Continental rise (1500-2200 m water depth; Sta. 6 and G).

Pore water profiles of Ge, Si, NH
4
+
, Fe
2+
, and Mn
2+
for a representative station of each type are given
in Fig. 4 (data is given in Table A2 of Appendix A). Pore water Si, NH
4
+
, Fe
2+
and Mn
2+

profiles for
several additional stations are shown in Fig. 5. The oxidants used to remineralize organic matter and the
resulting pore water profiles at each station are governed by the combination of factors, the chief being
the rate of biogenic matter rain to the seafloor and sediment accumulation, bottom water O
2

concentration, and benthic activity (bioturbation and bioirrigation). In terms of the sequence of electron
acceptors involved in organic matter (OM) remineralization, Type D stations show typical deep-water
profiles with very low NH
4
+
concentrations, Mn
2+
appearing only below 15 cm depth and no detectable

18

Fe
2+
within our sampling depth. Type C sediments become reducing within the first few centimeters, as
indicated by the appearance of Mn
2+
followed by Fe
2+
and NH
4
+
. Type A and B sediments are completely
Mn-reducing (as indicated by Mn
2+
benthic fluxes into the overlying water column; Berelson et. al, in
prep) and become Fe-reducing within the first centimeter. None of the sediments become sulfidic within
our sampling depth, as indicated by the lack of H
2
S odor and relatively constant total S concentrations of
28-32 mM at all stations, with a slight possible decrease with depth at Sta. 1, 4, and 10 (see Appendix A,
Fig. A1). The high rate of organic matter deposition and low-O
2
bottom water over the shelf means that
the O
2
and NO
3
-
become depleted within the top centimeter in Type A and B sediments. This results in a
situation where Fe- and Mn oxides may be the main electron acceptors, an unusual situation compared to
most global or coastal sediments but observed previously in similar deltaic settings (Aller et al., 1996;
Aller et al., 2004a; Aller et al., 2004b). Fe
2+
and Mn
2+
concentrations decrease to ~20 µM at almost all the
shelf stations below 20-30 cm depth (Figs. 4 and 5). Station 2 appears to be a somewhat special case, as
the sediments there go through an unconformity at 17 cm depth – what seems to be an upper layer of
calcareous ooze underlain by compacted mud below 25 cm.
The silicic acid pore water concentration trend among the stations (Type B > A > C ~ D) is similar to
the other solutes (Figs. 4 and 5). Except for Type A stations, the concentrations are on the low end of
observations in the marine environment (McManus et al., 1995; Gallinari et al., 2002). They range from
70 to 700 µM (Table 7), which is well below the bSi solubility of ~1 mM (e.g. Dixit et al., 2001). The
pore water profiles of several shelf stations do not reach a constant asymptotic concentration, instead
often decreasing with depth (especially pronounced at Type B stations). Type A stations have generally
lower and less variable Si concentrations (150-300 µM), while Type B exhibit higher and more irregular
Si concentrations in the range of 200-700 µM. Type C and D (continental slope and rise) stations have the
lowest pore water Si concentrations (60-200 µM).
Ge was only measured in pore waters from 4 cores (Fig. 4). Similar to Si, Ge concentration profiles
exhibit a near-surface maximum at each station and decrease further down, at similar depths. However,

19

the variations in Ge concentration are relatively much larger than those in Si concentration, as indicated
by a variable pore water (Ge/Si
PW
) ratio (Fig. 4). If Ge/Si
PW
was controlled solely by congruent bSi
dissolution, the slope of a plot of Ge vs. Si should stay constant with depth. This divergence of Ge and Si
concentrations is indicative of their differing behavior during sediment diagenesis and has been
previously observed in reducing California and Peru/Chile coastal margin sediments (Murnane et al.,
1989; Hammond et al., 2000; McManus et al., 2003), as well as in Southern Ocean (King et al., 2000).
We have observed multiple burrows in the sediment cores from most of the stations but made no
attempt to carefully quantify their frequency or depth range. A discussion of infaunal sediment mixing in
terms of the observed pore water profiles is given in Section 5.3.2.
4.3 Benthic fluxes
During core and benthic chamber incubations, the high concentrations of silicic and germanic acids in
pore waters of the upper sediments diffuse upwards, causing Si and Ge concentrations in the overlying
water to increase over time. These fluxes reflect dissolution of bSi, and perhaps other reactions.
Quantification of the diffusive (and possibly advective) fluxes provides estimates of net reaction rates.
Two to three cores per station were incubated, with good agreement between different cores at each
station (Fig. 6). The near-linear slopes suggest that the diagenetic reactions controlling the flux are
influenced by the same phase (or phases) during the whole course of the experiment. Only the first 2 time
points (0 – 15 h) were used to calculate the Si and Ge fluxes at Sta. M since cold room temperature could
not be kept close to in-situ for the remainder of the experiment.
Additionally, Si flux estimates were obtained from in-situ benthic chamber incubations (see
Methods). Values obtained by both methods are given in Table 2. Variability in Si flux as measured by
the standard deviation of the fluxes determined ranged from 1-33% for chambers and from 2-42% for
incubated cores. Thisvariability may be an effect of macrofauna mixing and irrigating sediment, which
results in spatial variability of various solute fluxes. It may also reflect patchy distribution of inputs of bSi

20

. No correlation between bottom water hypoxia and spatial variability was observed, suggesting that, if
macrofauna are responsible for flux variability, they are active throughout the continental shelf, as well as
in deeper waters. Both recovered core and in-situ incubations were carried out at Sta. 1 allowing for
comparison between the two methods. Core incubation seems to underestimate the Si flux by ~40% at
this station (Table 2) – possibly an effect of benthic fauna dying off as the overlying water becomes O
2
-
depleted over the course of the incubation. Similar discrepancy between these methods has been observed
previously, although to a lesser degree (Hammond et al., 2004a). However, the in-situ and core- estimated
Si fluxes are identical within 1σ S.D. of their replicates. The benthic chamber data was used in the
modelling of Ge and Si cycling at Sta. 1 (Section 5.2), as it is not influenced by temperature control
during the flux experiment and is thus likely to be more accurate.
The Ge vs Si concentration slope at each station represents the Ge/Si ratio of the flux (Ge/Si
flux
) into
the overlying bottom water (Fig. 6). Another way to estimate this value is from separately calculating the
flux of each element (Table 2). This latter approach results in slightly higher uncertainties but the values
obtained are within uncertainty indistinguishable from the slopes in Figure 6. Each of the three stations
exhibit distinct Ge/Si
flux
slopes, indicating that the Ge/Si ratio of the dissolving bSi varies among the
stations or that multiple phases may be reacting (dissolving, precipitating) in varying proportions. See the
discussion in Sections 5.2 and 5.3.4.
4.4 Sediments
Using alkaline leaching of bulk sediment we have determined the weight fraction of biogenic silica in
the 0-2 cm of sediments at Sta. 1 and Sta. G (Table 3). The results agree with previously reported values
(Turner and Rabalais, 1994; Presti and Michalopoulos, 2008). Using mass accumulation rates estimated
from
210
Pb measurements by other researchers we have calculated biogenic silica burial rates for both
stations (Table 3), assuming bSi does not change significantly below 2cm. Based on the pore water
profiles, it can be argued that bSi dissolution at Sta. 2 and Sta. M continues as far as 10 cm depth.

21

However, previous solid phase bSi analyses of Mississippi delta sediments have shown no appreciable
down core trends of %bSi within top 1-3 m of sediment (Presti and Michalopoulos, 2008).
The absolute Si burial flux is ~10 times higher at the shallow shelf Sta. 1 than at the deep offshore
Sta. G. However, bSi burial efficiency at each station was calculated as the ratio of the burial flux to the
opal rain to the sediments (the latter defined as the sum of diffusive and burial fluxes) and surprisingly is
similar, 10-20% for both sites (Table 7). Alkaline leach of Sta. 1 yielded lower %bSi but much higher
Ge/Si ratio than Sta. G (Table 3). Pretreating Sta. G sediments with HCl prior to alkaline leaching, yields
slightly higher %bSi (similar to the results of Presti and Michalopoulos (2008)) and about twice the Ge/Si
value. These results are discussed in Section 5.3.4.
Finally, we have purified and cleaned bSi from Sta. M sediments (see Methods), and measured it to
have Ge/Si
opal
of ~0.75 µmol/mol. However, this value should be taken cautiously, as it does not include
particles smaller than 20 µm and there were indications of Si polymerizing after bSi dissolution.
Therefore, our measured Ge/Si
opal
value for Sta. M should be taken as a lower estimate.
5. DISCUSSION
5.1 Water column and biogenic fractionation
The full shelf water column dataset yields a Ge/Si slope of 2.16 ± 0.08 µmol/mol (Fig. 3b), which is
significantly higher than either the global ocean value of 0.76 µmol/mol (Sutton et al., 2010) or the 1.56 ±
0.07 µmol/mol we have measured in the Mississippi river. The Mississippi-Atchafalaya river system is
the main source of nutrients (including Si) to the area, fueling diatom blooms that often result in hypoxic
waters over the shelf (e.g. Bianchi et al., 2010). However, most of the individual shelf water samples
exhibit Ge/Si
SW
ratios in the 2-3 µmol/mol range, with significant variability between different stations, as
well as elevated surface water Ge/Si
SW
at several stations (Figs. 2 and 3a). At a first glance, these
observations fit well with the recent research showing that (small) diatoms discriminate against Ge while
growing under very Si-limiting conditions (see Sections 1 and 4.1), which is likely for several stations

22

based on low surface Si concentration (Fig. 2). However, there are several other ways that such variability
could potentially be explained and those need to be addressed.
1) It has been shown that Ge/Si ratio of a particular river may change seasonally (Mortlock and
Froelich, 1987). Therefore shelf Ge/Si
SW
further from the river plume could reflect river water
delivered during the spring high discharge season, whereas stations closer to the plume would
reflect the current river signature. However, the high seasonal variation has been observed in
small Alaskan streams (Mortlock and Froelich, 1987), while the Mississippi River has one of the
biggest drainage basins in the world, encompassing ~40% of the conterminous United States.
Therefore, it should be largely insensitive to natural Ge/Si variations within different parts of the
basin.
2) A comprehensive survey by Froelich et al. (1985a) has shown that streams and rivers unaffected
by anthropogenic activities generally display Ge/Si ratios in the range of 0.3 – 1.2 µmol/mol,
while rivers draining industrialized regions generally exhibit higher Ge/Si of ~5 µmol/mol, most
likely due to contamination by coal and coal-derived fly ash. The Mississippi falls in the latter
category and Froelich et al. (1985a) have made multiple measurements, showing that average
Mississippi river Ge/Si was 3.7 ± 1 µmol/mol in the 1980s, relatively constant with changing Si
concentration. Interestingly, in 2011 we have measured a much lower Ge/Si ratio of 1.56
µmol/mol, which might be a result of the widespread implementation of coal ash capture and
storage in the recent decades. It is therefore also unlikely that the high shelf water Ge/Si
SW
ratios
are the result of direct fly ash deposition.
3) Previous studies of smaller river estuaries have shown the absence of significant Ge/Si
fractionation during freshwater and seawater mixing (Froelich et al., 1984; Froelich et al., 1985b).
Although the shelf waters have higher Ge/Si than the Mississippi river, a plot of shelf Ge/Si
SW
vs.
salinity shows no trends (data not shown). Similarly, there is no correlation of Ge/Si
SW
and

23

proximity to the river delta. Therefore, some process of Ge pre-concentration during freshwater-
seawater mixing can be ruled out as well.
Keeping all of the above in mind, the case can be made that Ge/Si
SW
enrichment occurs during
biogenic silica production in the water column, as diatoms discriminate against Ge during Si uptake, thus
elevating Ge/Si
SW
in the surrounding water. As mentioned previously, such a phenomenon has been
observed in Si depleted (<5 µM) surface waters of the ocean, resulting in Ge/Si
SW
ratios as high as 15
µmol/mol (Sutton et al., 2010; Froelich et al., 1989; Ellwood and Maher, 2003), while Ge/Si
diatom
remains
< 1 µmol/mol (Sutton et al., 2010).
While small diatoms seem to discriminate against Ge only while growing under very low Si
concentrations, large diatoms in Southern Ocean south of Antarctic Polar Front also fractionate Ge/Si,
despite growing in Si replete conditions (Shemesh et al., 1989). Sutton et al. (2010) have proposed that
Ge discrimination occurs only when diatoms engage in active Si uptake into the cell, using Si transporter
ligands to more efficiently transfer Si to the cell wall (Thamatrakoln and Hildebrand, 2008). One can
speculate that these transporter molecules would be more selective towards Si than Ge. On the other hand,
when growing in Si replete conditions, diatoms seem to rely more on diffusion-mediated uptake of Si
(Thamatrakoln and Hildebrand, 2008). Since germanic and silicic acids in water should have very similar
diffusivities (Hammond et al., 2000), such uptake should not result in significant Ge/Si fractionation.
However, Sta. 1 and Sta. 2 have similarly low surface water Si concentrations (2-3 µM) at the time of
our measurement but Ge/Si
SW
is much higher at Sta. 2 (Fig. 3). If diatoms discriminate against Ge during
Si uptake, then this could be explained by more complete Si utilization at Sta. 2. We can speculate why
this might be the case: 1) Sta. 2 is much further west from the river discharge (see Fig. 1), and thus likely
less turbid, aiding diatom productivity; 2) Sta. 2 is ~10 m shallower than Sta. 1, meaning that the
sediment-water interface receives more sunlight, aiding benthic diatom productivity and bSi burial in the
sediments. The degree of Si depletion or enrichment at any water depth can be quantified by comparing

24

the measured Si concentration to that predicted by conservative mixing of river and offshore GOM water
(Si
cons
) on a Si vs. salinity plot (solid line in Fig. 7A), an approach used previously in the same study area
by Nelson and Dortch (1996). The Mississippi River was chosen as the fresh-water end-member. For an
offshore water that advects onto the shelf, the water at 60 m depth at station G was chosen. While the
potential density of Sta. 1 bottom water is better matched by 40 m depth water at Sta. G,this was the
closest sample to 40 m available.
It is clear from Fig. 7a that Si cycling at these stations is dominated by biological effects, as all but
one sample fall off the conservative mixing line. The degree of Si drawdown or remineralization at each
depth is visualized in Fig. 7b. Si/Si
cons
= 0 represents 100% Si utilization, while Si/Si
cons
> 1 represents Si
enrichment due to bSi dissolution. Surface water Si utilization reaches 92 - 94% at Sta. 1 and 2 and 79%
at Sta. 9 (inset of Fig. 7b). However, bSi is remineralized to a much higher degree in the bottom waters of
both Sta. 1 and Sta. 9. Intense remineralization should act to resupply surface waters with Si and limit
biofractionation, keeping Ge/Si
SW
relatively low. At Sta. 2, on the other hand, Si uptake seems to be
outcompeting remineralization, resulting in very low Si concentration down to ~15 m depth and
extremely high Ge/Si
SW
ratios (Fig. 2). Although this correlation does not prove causation, it fits well
with previous observations of Ge/Si fractionation by various biosilicifiers (see above).
5.2 Ge/Si fractionation based on mass balance for shelf water and sediments
A simple steady-state box model of the shelf surface waters at Sta. 1 was built to calculate Ge/Si
fractionation associated with biogenic silica production in the surface water and the benthos required to
produce the observed Ge/Si
SW
enrichment (Fig. 8). Sta. 1 was chosen for the model due to abundant water
column data and its location in the Mississippi river plume, allowing us to make the approximation that
most of the Si and Ge are supplied to the box by the river.
The model is salt-balanced after correction for precipitation-evaporation (Table 5). The shelf surface
water box receives river water which mixes with the surface ocean water and also vertically with the

25

bottom water box (vertical wave mixing has been shown to be much more important that horizontal
mixing for the dilution of the Mississippi river plume (Wright and Coleman, 1971)) and gets advected
laterally to maintain mass balance. Similarly, the bottom water box receives an advective input of high
salinity offshore water. Water from 60 m depth at Sta. G (our furthest offshore station) was selected as the
offshore source by matching potential density with the bottom water of Sta. 1. The water and salt balance
equations (Eqs. 1-4) are given below, with variables shown in Fig. 8.
Water balance
Surface box: Q
r
+ Q
u
+ Q
pe
= Q
d
+ Q
off
(Eq. 1)
Bottom box: Q
d
+ Q
in
= Q
u
(Eq. 2)
Salt balance
Surface box: Q
u
S
b
= (Q
d
+ Q
off
)S
s
(Eq. 3)
Bottom box: Q
d
S
s
+ Q
in
S
d
= Q
u
S
b
(Eq. 4)
where Q
r
is the river water input, Q
pe
is the net precipitation-evaporation, Q
u
and Q
d
are the vertical
mixing fluxes between the surface and the bottom box, up and down, respectively. Q
off
is the water flux
out of the surface box, Q
in
is the advection of flux of offshore water into the bottom box. S
s
, S
b
, and S
d
is
the salinity of the surface, bottom, and advecting deep water, respectively.
Si and Ge cycles in the box model are controlled by the same set of processes: uptake in the surface
waters by planktonic diatoms, producing bSi flux J
p1
that sinks. A fraction, f
D
, of J
p1
is dissolved in the
bottom water box. The remainder reaches the seafloor. We adopt a value of f
D
= 0.5 ± 0.2 based on values
estimated by DeMaster et al. (1983) for the Amazon continental shelf. Although they had estimated 77-85
% dissolution of bSi in the water column, their budget did not account for additional Si burial as

26

authigenic silicates, which has since been shown to be prevalent in that environment (Michalopoulos and
Aller, 2004), significantly shifting the budget to better bSi preservation.
Additional Ge and Si is taken up by “bottom” silicifiers (e.g. sponges and radiolarians) and becomes
part of the sediment box, denoted as J
p2
. The fluxes of solid bSi to the sediment (1-f
D
) × J
p1
and J
p2

partially dissolve to sustain the benthic flux J
e
and the remainder is buried as J
b
. The equations describing
these processes are given below. Note that Eqs. 5-7 have both Si and Ge formulations.
Si & Ge balance
Surface box: Q
r
C
r
+ Q
u
C
b
= J
p1
+ C
𝑠 (Q
d
+ Q
off
)

(Eq. 5)
Bottom box: Q
d
C
s
+ Q
in
C
d
+ J
e
+ f
D
J
p1
= Q
u
C
b
+ J
p2
(Eq. 6)
Sediments: (1 - f
D
)J
p1
+ J
p2
= J
e
+ J
b
(Eq. 7)
In Eqs. 5-7, C is the concentration of either dissolved Ge or Si (subscripts same as in Eqs 1-4).
Values of C
r
, C
s
, C
b
, C
d
, S
s
, S
b
, and S
d
have been measured (Table 4). The diffusive flux out of
sediments J
e
of each element was determined from core incubation experiments at this site (Table 2).
Burial flux J
b
Si was determined from sediment mass accumulation rate and %bSi (Table 3).
This leaves a system of 10 equations that can be solved analytically for the 10 unknown values (Q
r
,
Q
u
, Q
d
, Q
in
, Q
off
, J
p1
Si, J
p2
Si, J
p1
Ge, J
p2
Ge, and J
b
Ge). The model allows the calculation of Ge burial flux
along with Ge/Si
opal
ratio in surface-formed, bottom-formed, and total buried bSi. Any additional Ge that
is buried then must be associated with the non-opal diagenetic sink. These and other results derived from
the model output are given in Table 6.
The model calculates Ge/Si
surface-opal
of 1.69 µmol/mol (Table 6) which includes all bSi
productivity in the upper 6 m of the water column at Sta. 1. This value is close to the Mississippi River
+ 0.05
- 0.03

27

Ge/Si of 1.56 ± 0.07 µmol/mol but significantly lower than either the surface water Ge/Si
SW
at Sta. 1
(3.15 µmol/mol, see Table 4) or the Ge/Si slope obtained from all our continental shelf water samples
(2.16 ± 0.08 µmol/mol, see Fig. 3b). Similarly, Ge/Si
bottom-opal
, which is inclusive of all bSi formed in the
water column below 10 m depth (most likely radiolarians and sponges), as well as the water-sediment
interface, is much lower (1.20 µmol/mol, Table 6) than either Ge/Si
surface-opal
or the bottom water
Ge/Si
SW
. The model does not include any assumptions of biological productivity and simply balances the
mass fluxes of Si and Ge. Therefore, the fact that the calculated Ge/Si
surface-opal
and Ge/Si
bottom-opal
values
are lower than the Ge/Si
SW
from which they are derived implies biological fractionation by the silicifying
organisms. A higher degree of biofractionation is observed in the bottom box, which is encouraging, since
both radiolarians (Shemesh et al., 1989) and sponges, especially, (Ellwood et al., 2006) are known to
discriminate against Ge strongly. Our model results offer further support for the hypothesis that surface-
water dwelling diatoms in the GOM do this as well. Similar fractionation factors K
D
are obtained for both
surface water dwelling diatoms and benthic silicifiers (presumably sponges). It must be noted that if Ge/Si
fractionation in silicifiers is indeed due to slightly different uptake kinetics of each element, then K
D

values observed will be dependent on the Si concentration in the seawater and the type of biosilicifying
organism. This relationship can be described using Michaelis-Menten uptake kinetics, where it is assumed
that organisms use the same uptake pathway for Si and Ge and any fractionation results from slightly
different uptake kinetics of each element (Ellwood et al., 2006), such that:
𝐺𝑒 /𝑆𝑖
𝑜𝑝𝑎𝑙 =
𝑉 𝐺𝑒
𝑉 𝑆𝑖
× 𝐺𝑒 /𝑆𝑖
𝑆𝑊
=
𝑉 𝐺𝑒
𝑚𝑎𝑥
[𝑆𝑖
′
]
𝐾𝑚
𝐺𝑒
+[𝑆𝑖
′
]
𝑉 𝑆𝑖
𝑚𝑎𝑥 [𝑆𝑖
′
]
𝐾𝑚
𝑆𝑖
+[𝑆𝑖
′
]
× 𝐺𝑒 /𝑆𝑖
𝑆𝑊
(Eq. 8)
where V is the uptake rate, V
max
is the maximum uptake rate, and Km is the Michaelis-Menten half
saturation constant of the respective element, and [Si’] is the sum of Si and Ge concentrations in the
seawater. Since [Si] >> [Ge], [Si’] can be approximated as [Si]. Further, we assume that the same
transport system is used for both elements, therefore V
Gemax
= V
Simax
and Eq. 8 can be expressed as
+ 0.11
- 0.14

28

𝐾𝑚
𝐺𝑒
= (
(𝐾𝑚
𝑆𝑖
+[𝑆𝑖 ])×
𝐺𝑒
𝑆𝑖
𝑆𝑊
𝐺𝑒
𝑆𝑖
𝑜𝑝𝑎𝑙 ) − [𝑆𝑖 ] (Eq. 9)
Nelson and Dortch (1996) performed kinetic Si uptake experiments using diatoms collected from the
Mississippi River plume in the summer of 1992 and measured a Km
Si
value of 5.3 ± 3.7 µM at [Si] = 4.7
µM at a location close to Sta. 1. Interestingly, this is identical to Cs
Si
used in our box model (Table 4) as
well as very similar to the threshold where Ge/Si fractionation is observed in the global ocean (Sutton et
al., 2010). We can use this value, Ge/Si
surface-opal
from Table 6 as Ge/Si
opal
, and surface seawater Si and
Ge/Si
SW
from Table 4 to calculate Km
Ge-surface
using Eq. 9 and obtain a value 13.9 µM. Although this
value is obtained very indirectly, to the best of our knowledge, there are no other Km
Ge
estimates for
diatoms at the moment. However, Sutton et al. (2010) used Km
Si
= 5 µM and either Km
Ge
= 9 or Km
Ge
=
12 µM to successfully model the global ocean Ge/Si
SW
distribution, suggesting that the biofractionation
calculated by our model is realistic.
We can use the same approach to model Ge/Si fractionation in the bottom box, using the equivalent
bottom box concentrations from Tables 4 and 6, and Km
Si
of 14.2 µM determined by Ellwood et al.
(2006) for sponges growing at various depths and Si concentrations (although none on continental
shelves). We obtain Km
Ge-bottom
= 48.5 µM, which is significantly lower than 173 µM calculated by
Ellwood et al. (2006) and confirms that Si uptake in the bottom may not be limited to sponges but most
likely includes some radiolarians and possibly diatoms. Indeed, our own visual observations, as well as
those of Jendrzejewski and Hart (1978) (albeit very qualitatively) show that diatom frustules (of which
~30-50% are from benthic species), sponge spicules, and radiolarian tests all occur with similar frequency
in the northern GOM shelf sediments. Diatoms are known to be more susceptible to dissolution than
radiolarians (Mortlock and Froelich, 1989) or sponges (e.g. Kamatani and Oku, 2000), and this is
reflected in the results of our model. Based on Ge/Si
opal
that needs to be buried to sustain steady-state
conditions in the water column, “surface” bSi (i.e. planktonic diatom frustules) contributes only 4-26 % of
total bSi burial (f
surface-buried
in Table 6), while the rest is comprised of bSi formed in the bottom box (note
+ 9.3
- 8.1

29

that this value is independent of the benthic J
e
flux – relevant to the discussion of different pore water Ge
sources below).
Preferential preservation of several well-silicified diatom species in the Indian Ocean continental
slope sediments has been previously observed (Koning et al., 1997). Although we do not have direct
evidence for this in the GOM, the fact that the more heavily silicified sponge spicules and radiolarians
have been observed with similar frequency to the generally smaller and less silicified diatoms supports
the contention that the former are selectively preserved and thus comprise a larger weight fraction of the
buried bSi, despite the latter being much more abundant in the water column. This is in line with the
recently revised global Si budget which acknowledges the possibility that sponge spicules may be
disproportionately important for Si burial in coastal margin sediments (Tréguer and De La Rocha, 2013)
and highlights the possible utility of bulk Ge/Si
opal
as a tracer of selective burial of different types of opal.
It must be noted that the high pore water Ge maximum at Sta. 1 (Fig. 4) indicates that there may be an
additional source of dissolved Ge to the pore waters that could be either terrigenous material (e.g. FeOx)
or previously authigenically sequestered Ge. If any of this “additional” Ge is escaping into the overlying
water column, it would mean that, in the case of terrigenous Ge source, our model underestimates the
non-opal sink. In the case of an authigenic Ge source, it reflects a non-steady state of the system (e.g. a
seasonal phenomenon related to the extensive bottom water hypoxia). See section 5.3.5 for a further
discussion of this possibility. Note that this complication does not affect the calculation of Ge/Si
opal
and
biofractionation in the water column.
The best estimate non-opal Ge flux at Sta. 1 calculated by the model is 10 pmol m
-2
d
-1
, which
comprises 2% of Ge buried in the sediments or 0.2% of total opal rain-delivered Ge (Table 6). The large
uncertainty range is due to the sensitivity of the model to certain input parameters but at the same time
demonstrates that these sediments could easily switch back and forth between acting as either Ge sink or a

30

source. A comparison with Sta. G, as well as sites in California Borderlands and the Southern Atlantic,
where non-opal Ge sequestration has been previously documented, is given in Section 5.4.
5.3 Sediment diagenesis
5.3.1 Organic matter
There are several features of the pore water profiles, especially of the continental shelf stations, that
are worthy of investigation. As a general trend, NH
4
+
, Fe
2+
and Mn
2+
pore water concentrations decrease
(Figs. 4 and 5) with increasing water depth (Table 1) presumably as a consequence of lower amount of
OM reaching the seafloor at the deeper sites. However, some shelf stations (Type B) have distinctly
higher pore water NH
4
+
and Si concentrations (Fig. 5) than others (Type A), despite having similar
sedimentation rates (Table 3) and water depth. Several factors may be responsible for this. First, it is
possible that Type B stations receive a much higher rain of biogenic matter (OM and bSi) than Type A
stations. We did not measure particle export rates. However, despite biological productivity over the
GOM shelf being the highest within the extended river plumes (Rabalais et al., 2002), there is no
correlation of sediment pore water concentrations with either the surface water salinity or the bottom
water hypoxia. For example, Sta. 4 and Sta. 9 are similarly affected by river water plumes and hypoxia
(Fig. 1), however, their pore water concentrations of NH
4
+
and Si are very different (Fig. 5). On the other
hand, Fe
2+
and Mn
2+
profiles are very similar at both types of stations (Fig. 5), pointing to controlling
factors other than OM remineralization. The continental margin setting and the discharge of the
Mississippi-Atchafalaya river system onto the shelf means that there is no shortage of terrigenous MnOx
and FeOx supply (Trefry and Presley, 1982). Therefore, post-depositional processes, such as diagenesis,
irrigation, and sediment mixing, are most likely to influence the observed pore water profiles of these
elements.

31

5.3.2 Sediment bioirrigation
Mixing and resuspension of sediments can occur either through wave currents (especially during
storm events), or through bioturbation and bioirrigation by benthic macrofauna, both of which are known
to occur in GOM continental shelf sediments (e.g. Allison et al., 2000; Corbett et al., 2006). These
processes act to 1) dilute the pore waters with overlying bottom water; 2) enhance and redistribute the
remineralization by mixing down fresh, more labile biogenic matter; and 3) affect redox-sensitive
elements, in cases where the bottom water and the pore water O
2
concentrations are different. For
example, NH
4
+
concentrations are expected to decrease with intense bioirrigation (Aller, 1980), while Si
concentrations have been modelled to either increase (Schink et al., 1975) or decrease (Aller, 1980),
depending on the approach and the assumptions in the models (e.g. the supply of labile bSi). In any case,
Si concentrations seem to be less sensitive to bioturbation than NH
4
+
.
In the pore water data presented here, Si and NH
4
+
profiles are similar to each other but very different
from either Fe
2+
or Mn
2+
at any given station (Fig. 5). If bioirrigation was the first order control of pore
water profiles, one would expect a similar shape profile for all the solutes. Furthermore, if bioirrigation
was significantly more intense at Type A relative to Type B stations (as may be invoked to explain the
large difference between Si (and NH
4
+
) concentrations at these sites), one would also expect the same
distinction for Mn
2+
and Fe
2+
, which is not the case (Fig. 5). Together, these facts suggest that benthic
infaunal activity does not exert the primary control on pore water concentrations.
5.3.3 Biogenic silica and reverse weathering
Excluding bioirrigation (see above), pore water Si may be chiefly controlled by the competition
between the rates of bSi deposition, dissolution, and re-precipitation of dissolved Si as authigenic
aluminosilicates, a.k.a. “reverse weathering” (see the review of Aller (2014) and references therein).
Water column productivity and depth controls the deposition (opal rain) rate, while the dissolution rate in
turn depends on the nature of bSi (diatoms vs radiolarians vs sponges) reaching the sediments and
alteration of this bSi in the water column and after burial (incorporation of Al and trace elements that may

32

“clog” reactive sites). Authigenic aluminosilicate precipitation has so far been observed in rapidly
accumulating sediments and seems to be limited by the supply of either bSi or lithogenic debris to the
sediments, as both components are essential (ibid.).
Although it is difficult to assess the rate of bSi delivery to the seafloor without sediment trap data,
benthic fluxes help estimate the amount of bSi dissolving in sediments. Averaging benthic flux data in
Table 2 yields only a ~40% higher average Si flux at Type B stations relative to Type A (6.8 and 4.9
mmol m
-2
d
-1
, respectively), while the difference in pore water concentration within the top 20 cm of
sediments (Table 3) is on the order of ~300%. Therefore, the large contrast of pore water Si
concentrations at the two stations does not seem to derive from the bSi deposition rate. Non-steady state
delivery of material to the sediments could also potentially explain the observed pattern. Benthic fluxes
most likely represent dissolution of relatively recently deposited material (e.g. bSi delivered during the
most recent diatom bloom, which could be similar for all shelf stations) while pore waters should reflect
deposition patterns of much longer timescales. Resuspension and focusing of sediments by wave action
could also act to obfuscate any long term patterns. However, it is impossible to assess this possibility with
the dataset presented here and we will assume it does not exert primary control over solute benthic fluxes
and pore water concentrations.
While most bioproductivity takes place at intermediate salinities within the Mississippi and
Atchafalaya river plumes (e.g. Pakulski et al., 2000), deposition and remineralization of the biogenic
matter happens over a much bigger area of the continental shelf, as indicated by the widespread hypoxia.
bSi dissolves more slowly in the water column than OM and therefore is likely to be dispersed even more
widely, as demonstrated, for example, in the sediments under the Amazon river plume (Chong et al.,
2014). This suggests that variability in bSi composition and lability is also unlikely to be varying
significantly among the shelf stations. The similar water depth of all the continental shelf stations also
rules out the possibility of different degrees of bSi alteration (e.g. Al incorporation) while it settles.

33

Having effectively ruled out all the other factors, we can try to use the pore water concentrations of
the different solutes to investigate the different diagenetic processes within the sediments. Most of the
suspended load from the Mississippi and Atchafalaya rivers settles out within ~30 km of the deltas but
some of it is often resuspended and spread further along the shelf, especially during high-energy winter
months (Allison et al., 2000; Corbett et al., 2007). Deeper water sediments accumulate ~10-40 times
slower but also receive ~10 times lower opal rain (compare Sta. 1 and Sta. G in Table 7). Type A stations,
being situated closer to either the Mississippi (Sta. 1) or the Atchafalaya (Sta. 4 and A) river mouth, are
likely to receive a higher flux of terrigenous river-borne material than Type B stations. A high detrital:bSi
ratio in the sediments is known to limit pore water Si concentrations to levels much lower than those
expected at thermodynamic equilibrium (McManus et al., 1995; Dixit et al., 2001; Khalil et al., 2007) due
to enhanced Al incorporation and conversion of bSi to authigenic silicates, i.e. “reverse weathering”,
which seems to be especially rapid in deltaic sediments (Michalopoulos and Aller, 1995; Michalopoulos
et al., 2000; Michalopoulos and Aller, 2004; Presti and Michalopoulos, 2008; Loucaides et al., 2010). The
combination of high bSi supply and intense Mn- and Fe- redox cycling in such environments seems to
enhance the process of reverse weathering (Aller, 2004). Indeed, Presti and Michalopoulos (2008) have
shown that up to 40% of biogenic silica in the Mississippi river delta sediments is converted to authigenic
secondary clay minerals.
Based on our pore water and bSi analyses, we suggest that this process occurs not only in the
immediate vicinity of the Mississippi River delta but also in the sediments of the distal shelf. For
example, Sta. 1 sediments are accumulating ~20× faster (Table 3) and receiving a much higher bSi flux to
the seafloor than Sta. G. However, pore water Si concentrations are similar at both stations (Fig. 4). While
opal preservation efficiency (the fraction of bSi rain reaching the sediments that is permanently buried)
seems to be actually higher at Sta. G (Table 7), this value is based on alkaline-leach determined %bSi,
which does not take into account the majority of authigenic aluminosilicates. It is therefore likely that the
calculated Si preservation would be much higher at Sta. 1, if reverse weathering products were taken into

34

account. Alteration and preservation of bSi might also play a role, especially in the deeper water Type C
and D stations since a deeper water column both eliminates the least-silicified frustules and allows more
time for possible Al incorporation during deposition.
Below a sediment depth of 20 cm, the Mn
2+
and Fe
2+
concentrations at all the continental shelf
stations converge towards a constant low value of ~20 µM (Fig. 5). Ruling out sediment bioirrigation (see
Section 5.3.2), this might be another indication of reverse weathering, since these elements are known to
be scavenged by authigenic aluminosilicates (Loucaides et al., 2010). However, the precipitation of other
authigenic minerals, such as carbonates or sulfides, could also be responsible (see Section 5.3.4). Finally,
despite generally noisy profiles, major cation concentrations decrease slightly with depth at most of the
shelf stations. This effect is especially pronounced for K
+
(Appendix A, Fig. A2) and less notable for Na
+

and Mg
2+
(data not shown). Once again, this is suggestive of reverse weathering reactions, since
authigenic clays forming in a similar setting in the Amazon delta (Michalopoulos and Aller, 2004), as
well as in incubation experiments performed with Mississippi river delta sediments (Loucaides et al.,
2010), have been shown to incorporate all of these elements (with a possible preference for K
+
), at the
same time reducing bSi solubility.
5.3.4 Non-opal Ge sink
If the dissolution of bSi were the sole process controlling the buildup of dissolved Ge and Si in the
pore waters, then Ge/Si
PW
should approach Ge/Si
opal
. Instead, Ge/Si
PW
varies significantly with depth at all
of the stations (Fig. 4), indicating either 1) diagenetic processes that affect Ge and Si to a different degree,
i.e. the dissolution and precipitation of phases other than bSi, with their own Ge/Si ratios, or 2)
bioirrigation, which could act to mix the pore waters and overlying bottom water to a different extent at
different depths in the sediment. A simple way to test the latter possibility is to plot Ge/Si
PW
against 1/Si
(Fig. 9). If Ge/Si
PW
were controlled solely by bioirrigation, the values should fall on a mixing line
between overlying bottom water and bSi. However, Sta. 2, M, and G all fall below the mixing line,
indicating preferential uptake of Ge into an authigenic phase forming in these sediments. Fig. 9

35

unequivocally shows that bioirrigation coupled with bSi dissolution is not (directly) responsible for the
observed Ge/Si
PW
variations, and a precipitation of a Ge-enriched phase is required for these three
stations. Sta. 1 seems to indicate the opposite trend – release of non-opal Ge (or preferential uptake of Si
over Ge, although this seems very unlikely). The possible causes and implications of this phenomenon are
discussed in Section 5.3.5.
A different way to estimate the extent of diagenesis is by measuring the benthic flux of each element
from sediments into the overlying water column (Fig. 6). Our results indicate similar or slightly higher
Ge/Si
flux
compared to what is buried at Sta. M and G. A higher benthic flux Ge/Si relative to what is
buried in the sediments very likely suggests the selective preservation of certain, better-silicified forms of
bSi, such as radiolarian tests or sponge spicules rather than diatom frustules. Another possibility is that Si
is being preferentially retained in the solid phase relative to Ge, however this is in clear disagreement with
the pore water results, which also generally display lower Ge/Si
PW
(except one horizon at Sta. 1)
indicating preferential Ge retention in the solid phase (see discussion below). As discussed previously,
modeling results at Sta. 1 suggest much better preservation of “bottom” bSi (see Section 5.2), which tends
to have lower Ge/Si
opal
due to a higher degree of biofractionation, while most of the more labile diatom
frustules would dissolve on top of the sediments and never get buried. They are also more likely to be
converted into secondary clays due to a larger surface area and a more delicate structure, compared to
radiolarian tests or sponge spicules. Therefore, the Ge/Si
flux
from core incubations most likely reflects a
larger contribution from planktonic diatoms, while Ge/Si
opal
measured in the sediments reflects a larger
contribution of radiolarians or sponge spicules to the buried bSi.
While Sta. 1 has a distinct high-Ge/Si source, which complicates the interpretation of the Ge/Si
flux

value, Sta. M and Sta. G show no similar indications, suggesting that bSi is the dominant source of Ge
and Si to the pore waters. If we assume that the rain reaching the sediments at the stations has Ge/Si = 1.1
µmol/mol (based on a leach of Sta. G sediments, see Table 3), and using the Ge/Si
flux
measured at each
station (Fig. 6) we can create a sedimentary Ge budget and calculate non-opal Ge loss or release from the

36

sediments (Table 7). A box model involving water column processes was used to achieve the same for
Sta. 1 (see Section 5.2).
Identifying the exact authigenic mineral phase responsible for Ge scavenging during marine sediment
diagenesis would be a big step towards constraining the modern Ge biogeochemical cycle, which would
then enable a much-better informed interpretation of Ge/Si
diatom
paleorecords (see Introduction). Ge and
Fe concentrations peak and then decrease downcore at all of the anoxic stations (Fig. 4), pointing to a link
between Ge and reduced Fe in marine sediments, in agreement with previous observations (Hammond et
al., 2000; King et al., 2000). A degree of decoupling between the pore water concentrations of the Si, Ge,
and Fe, however, results from the differences in their sources (bSi dissolution for Ge and Si, FeOx
reduction for Fe
2+
and possibly Ge (see Section 5.3.5)) and their relative partitioning into the authigenic
products, which are expected to have higher Fe/Si and Ge/Si ratios than bSi.
While some lab experiments have shown that Ge is incorporated into FeOx preferentially to Si
(Anders et al., 2003; Pokrovsky et al., 2006), field data from soils (Kurtz et al., 2002), glacial streams
(Chillrud et al., 1994), and hydrothermal plumes (Mortlock and Froelich, 1986; Mortlock et al., 1993)
suggest that this is not always the case and that this process might be very sensitive to the solution
composition or perhaps certain environmental variables. In fact, Tribovillard et al. (2011) have found that
in various geological formations dating as far back as the Cretaceous, Ge tends to be enriched in those
that were likely deposited under reducing conditions. From the pore water data presented here, it is
unclear if any Ge sequestration occurs at the Fe redox horizon. The decrease of Ge/Si
PW
in the top 3 cm of
sediments at Sta. 1 and Sta. 2 correlates with the very shallow Fe-redox zone (unresolved within our
sampling frequency) but no such correlation is observed for Sta. M, which has a well resolved Fe-redox
zone at ~6 cm depth. Nevertheless, previous studies of marine sediments have clearly shown that Ge is
not preferentially incorporated into FeOx in these environments (Murnane et al., 1989; Hammond et al.,
2000).

37

Certain sulfide minerals are also enriched in Ge (Bernstein, 1985). Sulfate reduction and iron sulfide
formation has been previously observed across the whole Mississippi delta and Texas-Louisiana shelf in
the general location of our study (Lin and Morse, 1991) and could present an alternative authigenic
silicate as the non-opal Ge sink. In one particular case, Lin and Morse (1991) have measured ~40% pore
water SO
4
2-
depletion at 12 cm depth of the sediments at a location 35km away from our station Sta. 4 and
~100% depletion by 70 cm depth. In contrast, we see very little change in total pore water S
concentrations with depth (Appendix A, Fig. A1) and could not detect the odor of H
2
S while sectioning
the cores, which suggests that the sulfidic zone in these sediments has shifted deeper since Lin and
Morse's (1991) study. A possible explanation could be that the upper ~1 m of sediments has been more
thoroughly mixed in the past two decades, for example, due to stronger or more frequent storm events in
the GOM. It is also possible that the sulfate reduction zone shifts up and down seasonally due to varying
bottom water O
2
content and benthic macrofaunal activity, although our samples were collected during
maximum seasonal hypoxia. It is unclear during what season and water column conditions Lin and Morse
(1991) collected their samples. Either way, pyrite formation most likely takes place in deeper sediments,
and some sulfide might even be forming within our sampling depth – either from S
2-
released during OM
remineralization, or through SO
4
2-
reduction that is small enough to not significantly affect the SO
4
2-
pore
water concentration. Indeed, small amounts of framboidal pyrite have been detected in authigenically
altered frustules from the Mississippi delta sediments (Presti and Michalopoulos, 2008). This means that
we cannot rule out Fe sulfides as a potential non-opal Ge sink in the GOM sediments. However, King et
al. (2000) have observed a strong Ge non-opal sink in South Atlantic sediments, where sulfate reduction
has been shown to be negligible down to ~100 m depth.
Ge is known to be preferentially incorporated into secondary clays during silicate weathering on land
(Froelich et al., 1992; Murnane and Stallard, 1990; Kurtz et al., 2002) suggesting similar partitioning may
take place during secondary clay formation in the marine environment. Another clue that reverse
weathering may be the process responsible for Ge sequestration is the fact that both the rate of non-opal

38

Ge retention (King et al., 2000; Esther et al., 2010) and the rate of authigenic aluminosilicate formation
(Loucaides et al., 2011) seem to increase with a higher detrital:bSi ratio in the sediments. Finally,
preliminary results from lab experiments also indicate preferential Ge incorporation into authigenic
silicates forming under anoxic, Fe-rich conditions (Tosca et al., 2014).
Preliminary alkaline leaching of Sta. 1 sediments (0-2 cm depth) released Ge/Si ~4.5 µmol/mol into
the solution (Table 3), supporting the hypothesis that Ge is enriched in a silicate phase, rather than FeOx
or FeSx, as neither of the latter two should have been leached significantly by the 5% Na2CO3 solution.
In addition, results from Sta. G (2100 m water depth) indicate that reverse weathering may take place
even in very slowly accumulating deep water sediments (Table 3). A small but significant decrease of Ge
and Si concentrations with depth is detected, accompanied by the decline of Ge/Si
PW
to values ~2-3 times
lower than Ge/Si
opal
(Fig. 4 and Table 7). Although dissolved Fe
2+
was undetectable within the top 37 cm,
it is very likely that the sediments become Fe-reducing further down.
Previous studies indicate phases other than FeOx or FeSx can sequester Ge in marine sediments. We
have presented evidence for the coincidental occurrence Ge sequestration and the formation of authigenic
aluminosilicate in these GOM sediments.
5.3.5 Non-opal Ge source
Although we have shown that bioirrigation cannot explain the downcore variations of pore water Si,
Ge, and Ge/Si, it could, however, have an indirect effect on the rate of diagenesis, for example, by
preventing the buildup of Si concentrations (thus enhancing bSi dissolution) or by mixing down FeOx,
which would ensure a constant supply of fresh substrate for bacterial dissimilatory Fe reduction (Burdige,
1993). The latter phenomenon seems especially likely for Sta. 1 which shows a clear enrichment of
Ge/Si
PW
(Fig. 4) that requires a source higher in Ge than bSi (Fig. 9). This possibility is further supported
by the disagreement between model-calculated sedimentary Ge/Si
opal
(the model assumes bSi is the only
Ge source to the sediments) and that obtained by leaching (Table 3). FeOx are a likely candidate, since

39

they’ve been shown to be enriched in Ge relative to many other crustal materials, and especially relative
to biogenic silica (Bernstein, 1985; Bernstein and Waychunas, 1987; Kurtz et al., 2002; Pokrovsky et al.,
2006; Mortlock et al., 1993). Indeed, Murnane et al. (1989) have observed a very similar phenomenon at a
site in the sediments of San Pedro basin in the Southern California margin. The fact that we observe this
at Sta. 1 but not the other stations could be a combination of two factors: 1) this station is relatively close
to the Mississippi river delta and is therefore more likely to receive a significant flux of river-borne
detrital material, including MnOx and FeOx; and 2) it is located right at the “edge” of the hypoxic zone at
the time of our sampling (Fig. 1), thus the sediments might experience dynamically changing OM rain
and/or bottom water O
2
concentration, which would in turn affect the availability of different oxidants in
the sediments. Since dissimilatory iron reduction is known to be carried out by facultative anaerobes
(Burdige, 1993), it is entirely possible that our pore water data are reflecting a period of rapid seasonal Fe
reduction as the bacterial community turns to oxidants other than O
2
. High pore water Mn
2+
concentration
in the top centimeter of sediments (Fig. 5) suggests that MnOx are reduced very quickly and might not be
available to the bacteria dwelling deeper.
Another possible source is the dissolution of previously formed authigenic aluminosilicates. It is
unknown how stable these minerals are under various redox or saturation conditions but they usually
incorporate iron as reduced Fe
2+
(Michalopoulos and Aller, 1995) and therefore are unlikely to serve as
biologically relevant electron acceptors.
Finally, it is possible that the pore water Ge maximum reflects an anthropogenic contamination event,
e.g. coal waste discharge into the Mississippi River, which would have caused high Ge concentrations in
the water column, which then through silicifier uptake, deposition, and dissolution could be transferred
into pore waters. However, if this were the case, sediment mixing and diffusion should act to ”smear” this
peak over a much wider depth range than observed. A higher resolution sampling of the pore waters
would be required to test this hypothesis. In addition, there are no similar Ge/Si
PW
peaks at any of the
other stations, reducing the likelihood of this possibility further.

40

Despite the source, this “extra” Ge released into pore waters is then likely to be incorporated in the
precipitating secondary aluminosilicates (as discussed above), enriching them to much higher Ge/Si
values than, for example, at Sta. G. Alkaline leaching of Sta. 1 sediments supports this contention. The
diffusive flux of Ge across the sediment-water interface at this station therefore represents a net balance
of two sources: bSi dissolution (Ge/Si
opal
= ~1.2 µmol/mol (see Section 5.2) and possibly FeOx reduction
(Ge/Si very high); and one sink: authigenic aluminosilicate precipitation (although re-oxidation of Fe
2+

might reclaim some Ge, it is likely that over long term Ge will still end up in the secondary clays as they
are unlikely to be as sensitive to redox changes caused by bioturbation or anaerobic respiration as Fe. The
good agreement between the model-calculated Ge/Si
opal
and the Ge/Si
flux
across the sediment-water
interface suggests that most of Ge released from Fe reduction gets captured and does not escape into the
water column. However, it is possible that Ge/Si
flux
in a way reflects the ratio of Fe reduction relative to
authigenic precipitation of aluminosilicates, which would be much higher at this station than, for
example, at Sta. G, where sediments accumulate very slowly and in the absence of significant Fe
reduction within the top 30 cm, therefore allowing more time for authigenic clays to form without the
buildup of high Fe
2+
and Ge concentrations.
5.4 Spatial extent and variation of the non-opal Ge sink
In order to examine the factors that might be controlling the rate and extent of non-opal Ge
precipitation, we have compared the parameters of Si and Ge cycling at our sites to those previously
described in the literature (Table 7). A range of Si flux values measured over different seasons is given
for the California Borderlands sites with the aim of highlighting the non-steady state nature of Si cycling
over such timescales. Highly variable benthic Ge/Si
flux
ratios have been observed in these sites, suggesting
oscillation between active non-opal Ge sequestration, no sequestration, and release of non-opal Ge
(represented by the range of f-ratios in brackets). For example, Hammond et al. (2000) and McManus et
al. (2003) have measured that ~50% of Ge released from bSi dissolution gets removed from the solution
in the Santa Monica Basin, while King et al. (2000) observed no Ge sequestration. Subsequently, we have

41

recovered and incubated sediment cores from this basin to measure their Ge/Si
flux
ratios. Despite being
collected on the same day and from the same site, each of the five cores have yielded a significantly
different Ge/Si
flux
, that also varied throughout the duration of the experiment (not shown). One way to
explain this data would be to invoke a second source (for example, FeOx or authigenic clay dissolution),
supplying Ge at a different rate than bSi dissolution. The abundance of this phase must vary among the
cores collected.
Table 7 illustrates that non-opal Ge sequestration occurs over a wide range of depositional
environments. Non-opal Ge burial fluxes in the GOM sediments are about an order of magnitude lower
than in the California continental margin basins and similar to the Southern Atlantic sediments. There
does not appear to be correlation between any of the bSi fluxes and the Ge non-opal parameters. In the
South Atlantic sites, the detrital:opal ratio of the sediments has been shown to correlate with the f-ratio
(King et al., 2000) but detrital:opal ratios are unavailable for the other sites given in Table 7 to see if the
relationship holds true. However, it seems that the most efficient non-opal Ge sequestration occurs in the
areas with the lowest Si burial fluxes, which might be indicative of a high degree of bSi conversion to
authigenic silicates.
6. CONCLUSIONS
1. Water column Ge/Si ratios over the northern Gulf of Mexico continental shelf are significantly
higher than in the global ocean due to the supply of Ge-enriched Mississippi-Atchafalaya river
water into the area and the discrimination against Ge by planktonic and benthic biosilicifiers,
especially at Si < 5 µM. It is possible that the degree of seawater Ge/Si enrichment correlates
with the balance of Si drawdown vs. remineralization at a given station.
2. Steady-state box modelling of continental shelf water and sediments requires Ge/Si fractionation
during Si uptake to balance the water column and benthic Ge/Si budgets. The effect for benthic
silicifiers, assumed to be radiolarians and sponges, indicates they fractionate Ge/Si more strongly
and are also more important to biogenic silica accumulation and burial in these sediments than the

42

planktonic diatoms. Comparison of different silicifier Ge/Si signatures with what is buried in the
sediments could be useful in determining the contribution of these different organisms to the
sedimentary bSi budget at a given environment.
3. Based on pore water and bSi analyses, we propose that authigenic aluminosilicate precipitation
and alteration of biogenic silica (reverse weathering) is prevalent not only in deltaic sediments, as
reported previously (Presti and Michalopoulos, 2008), but throughout the continental shelf, slope,
and rise sediments of northern Gulf of Mexico. This process also likely incorporates Ge at a
higher Ge/Si ratio than the dissolving bSi, contributing to the “non-opal Ge sink”, although the
contributions of other phases (especially pyrite) cannot be ruled out completely. More thorough
investigation of the solids is required to assess the fraction of Si and Ge sequestered in this phase.
4. A summary of previously published Si and Ge fluxes at several locations in the ocean indicates
that this non-opal Ge sink may be more variable spatially and or temporally than previously
thought, although the factors responsible for this variability are unclear.

43

7. TABLES AND FIGURES
Table 1. Sampling station details
Station Casts
a
Data
b

Sampling
date
Latitude,
dec. deg.
Longitude,
dec. deg.
Bottom
depth,
m
Bottom
O
2
,
d

µmol/kg
1 CTD- 5/6, MC-2/3
WC, PW, CI,
BC, SP
2011-08-01 28.591 -90.541 30-38 31-51
A CTD-8, MC- 4 PW 2011-08-02 28.759 -92.184 40 155
2 CTD-10, MC-5/6/7 WC, PW, CI 2011-08-03 29.198 -92.919 22-24 27
3 CTD-18, MC-9 PW, BC 2011-08-04 29.115 -93.224 23 153
4 CTD-20, MC-10 PW, BC 2011-08-05 29.147 -91.942 17 4
5
CTD-23, MC-
11/13
WC, PW, CI 2011-08-06 28.199 -92.081 90 143
6 MC-14 PW, BC 2011-08-08 26.963 -92.285 1547 174
G
CTD-32/33, MC-
15
WC, PW, CI,
SP
2011-08-09 26.372 -92.141 2141 209
7 CTD-36, MC-17 WC, PW, BC 2011-08-12 27.952 -92.128 145 116
M CTD-41, MC-18 PW, CI, SP 2011-08-13 27.911 -91.797 438 115
9 CTD-45, MC-22 WC, PW, BC 2011-08-15 28.971 -90.402 22 11
10 CTD-46, MC-26 PW, BC 2011-08-17 30.189 -88.629 23 128
MR
c
bucket WC 2011-08-20 29.934 -90.050 - -
a
Listed only those stations data from which is presented in this study. In the text, where data from multiple casts at one station is
presented the specific cast will be noted, otherwise, only the station number will be referred to.
b
WC = water column (referring to Ge and Si conc.); PW = pore waters; CI = core incubations; BC = benthic chamber
incubations; SP = solid phase data (chemically leached or physically separated bSi).
c
Mississippi River, New Orleans.
d
The hypoxic zone was dynamically shifting during our cruise, therefore these values may not be representative of the prevailing
conditions over the months prior. See Fig. 1 for a more robust mapping of the hypoxic zone near the time of our cruise by
(Rabalais and Turner, 2011).

44

Table 2. Benthic Si and Ge fluxes

Benthic chambers Core incubations
Station Chamber
Si flux, mmol
m
-2
d
-1

Core
Si flux, mmol
m
-2
d
-1

Ge flux, nmol
m
-2
d
-1

Ge/Si flux*,
µmol/mol

Sta. 1

4.11 ± 1.37

2.43 ± 0.67 2.22 ± 0.08 1.19 ± 0.28

Yellow 2.7 MC-2A 2.09 2.17 1.04

Red 4.18 MC-2B 1.65 2.28 1.38

Blue 5.44 MC-3A 2.88

MC-3B 3.08

Sta. 2

1.86 ± 0.19

MC-6A 1.97

MC-6B 1.98

MC-7A 1.90

MC-7B 1.58

Sta. 3

5.29 ± 0.76

Red 4.75

Blue 5.82

Sta. 4

8.16 ± 3.02

Yellow 5.43

Red 7.63

Blue 11.41

Sta. 5

0.64 ± 0.27

MC-11A 0.63

MC-11B 0.54

MC-11C 0.38

MC-11D 1.01

Sta. 6 Blue 1.56

Sta. 7 Red 0.80

Sta. 9

11.45 ± 2.72

Yellow 8.58

Red 11.77

Blue 14

Sta. 10A

7.70 ± 0.06

Red 7.74

Blue 7.65

Sta. 10B

9.16 ± 0.45

Yellow 9.55

Red 9.26

Blue 8.67

Sta. M

0.59 ± 0.01 0.64 ± 0.15 1.09 ± 0.24

MC-18A 0.59 0.79 1.34

MC-18B 0.57 0.64 1.11

MC-18C 0.60 0.48 0.81
Sta. G

0.18 ± 0.04 0.15 ± 0.05 0.87 ± 0.38

MC-15A 0.16 0.12 0.75

MC-15B 0.22 0.21 0.95

MC-15C 0.15 0.13 0.87
Uncertainty of station averages given as ± 1σ S.D., where multiple cores or chambers were used.
* Calculated as the ratio of average fluxes only from cores where both Si and Ge data are available. Uncertainty propagated from
the corresponding flux uncertainties.

45

Table 3. Summary of sediment results and characteristics

Sta. 1 (TypeA) Sta. 2 (Type B) Sta. M (Type C) Sta. G (Type D)
Biogenic silica
a
Ge/Si, µmol/mol 4.5 (1.2 - 1.3)
d
- 0.75
i
0.6 (1.1)
j

%bSi 0.53 ± 0.06 % - - 1.24 (1.38) ± 0.07 %
j

Benthic flux
b

Si, mmol m
-2
d
-1

2.43 ± 0.67 (4.11 ±
1.37)
1.86 ± 0.19 0.59 ± 0.01 0.18 ± 0.04
Ge, nmol m
-2
d
-1
2.22 ± 0.08 - 0.64 ± 0.15 0.15 ± 0.05
Ge/Si, µmol/mol 1.24 ± 0.13 - 1.04 ± 0.04 0.88 ± 0.11
Burial flux
c
Si, mmol m
-2
d
-1
0.44 - - 0.045
Pore water
Ge/Si, µmol/mol 1.0 - 4.3 0.6 - 1.2 0.4 - 0.6 0.3 - 0.6
Ge, pM (sampling
depth)
200 – 740 (0 – 15 cm)
330 – 640 (0 – 20
cm)
53 – 83 (0 – 39 cm) 52 – 64 (0 – 39 cm)
Si, µM (sampling depth) 130 – 210 (0 – 15 cm)
160 – 580 (0 – 33
cm)
100 – 170 (0 – 39
cm)
90 – 170 (0 – 39 cm)
Sedim. rate,
cm yr
-1
(g cm
-2
yr
-1
)
0.2 (0.18)
f,g
0.4
h
0.03 - 0.06 (0.17)
e
0.01 (0.008)
e

Bottom water
O 2, µM
31 – 51 27 115 209
a
Measured in top 2 cm of sediments, using alkaline leach; %bSi defined as wt% SiO 2 (see Methods).

b
Fluxes measured in core incubations, except for the value in brackets for Sta. 1, which is benthic lander data. Benthic flux Ge/Si ratio from
slopes in Fig. 6.
c
Calculated from sediment mass accumulation rate and %bSi.
d
Value in brackets is calculated by the model (Table 6). The disagreement between the measured and the calculated values is discussed in
Section 5.3.4.
e, f, g, h
Taken from
210
Pb-based estimates of
e
(Gordon and Goñi, 2004),
f
(Allison et al., 2000),
g
(Corbett et al., 2006),
h
(Turner et al., 2004).
i
Measured in bSi purified from 2-4 cm sediment depth, a lower-end estimate (see Methods).
j
Values in brackets for Sta. G are the results of alkaline leach after acid pre-treatment, similar to (Presti and Michalopoulos, 2008), see Methods.

46

Table 4. Water column data from Sta. 1 used in the box model
Station
Role in
model
Sample
name
Depth, m Ge, pM Si, µM
Ge/Si,
µmol/mol
Salinity, psu
MR River input - 0
C
r
Ge=264 C
r
Si=169
1.56 0
G
Deep water
input
CTD-33-11 60
*
C
d
Ge=3.3 C
d
Si=1.9
1.75
S
d
=36.63
1

CTD-5-12 1 15.5 3.2 4.93 26.1

CTD-5-11 4 9.3 3.2 2.86 26.6

CTD-5-10 6 19.6 7.7 2.55 27.3

Surface
water box
Average

C
s
Ge=14.8 C
s
Si=4.7
3.15
S
s
=26.67

CTD-5-9 10 28.9 14.1 2.05 34.0

CTD-5-8 12 42.7 20.2 2.11 34.8

CTD-5-6 16 83.7 37.3 2.24 35.5

CTD-5-5 18 80.4 37.7 2.14 35.7

CTD-5-4 20 75.6 36.8 2.06 35.8

CTD-5-3 22 72.9 35.8 2.04 35.8

CTD-5-2 24 70.0 30.8 2.27 35.9

CTD-5-1 24 66.6 30.9 2.16 36.0

Bottom
water box
Average

C
b
Ge=64.5 C
b
Si=30.2
2.13
S
b
=35.33
*
Determined this depth as the source of advecting water by matching potential density to the bottom waters of Sta. 1 (see Section 5.1).

47

Table 5. Box model input parameters and ranges tested
Parameter Main input Range tested
Precipitation-Evaporation (Q
pe
)
a
1.92
b
mm d
-1
±50%
bSi fraction dissolved (f
D
) 0.5 ±0.2
Si flux from sed. (J
e
Si) 4.11
c
mmol m
2
d
-1
±1.37
Ge/Si flux from sed. (J
e
Ge/J
e
Si) 1.24 µmol/mol ±0.1
Si burial flux (J
b
Si) 0.44
d
mmol m
2
d
-1

+ 100%
- 25%
River Ge/Si (C
r
Ge/C
r
Si)
a
1.56 µmol/mol ±0.1
SW Si conc. (C
s
Si, C
b
Si, C
r
Si,C
d
Si) see Table 4 ±5%
SW Ge conc. (C
s
Ge, C
b
Ge, C
d
Ge)
a
see Table 4 ±5%
SW Salinity (S
s
, S
b
, S
d
) see Table 4 ±1%
a
The model is significantly more sensitive to these parameters, the uncertainty of which (i.e. the tested range) is the cause for most of the
uncertainty in model output (see Table 7).
b
Avg. 2011 July data, obtained from NASA TRMM Microwave Imager (TMI) and NOAA Multiple-Satellite Blended Sea Surface Winds /
Monthly / Wind Speed (NOMADS) in May 2013.
c
Benthic lander data (Table 2).
d
See Table 3. Uncertainty range was selected taking into account that our leaching method might underestimate burial of diagenetically altered
silica (Presti and Michalopoulos, 2008).

48

Table 6. Box model results
Parameter Best estimate
a
Range
b
Expression
Ge/Si
surface
, µmol/mol 1.69
+ 0.05
- 0.04
J
p1
Ge/J
p1
Si
Ge/Si
bottom
, µmol/mol 1.20
+ 0.11
- 0.14
J
p2
Ge/J
p2
Si
Ge/Si
opal
, µmol/mol 1.24
+ 0.07
- 0.05
(J
p1
Ge*(1 - f
D
) + J
p2
Ge)/(J
p1
Si*(1 - f
D
) + J
p2
Si)
f
surface-buried
c
9 %
+ 17 %
- 5 %
(Ge/Si
opal
- Ge/Si
bottom
) / (Ge/Si
surface
- Ge/Si
bottom
)
K
D-surface
d
0.54
+ 0.07
- 0.06
(J
p1
Ge/J
p1
Si)/(C
s
Ge/C
s
Si)
K
D-bottom
d
0.56
+ 0.04
- 0.03
(J
p2
Ge/J
p2
Si)/(C
b
Ge/C
b
Si)
Km
Ge-surface
e
13.9
+ 9.3
- 8.1
Eq. 9
Km
Ge-bottom
e
48.5
+ 4.6
- 1.9
Eq. 9
Ge opal burial flux,
pmol m
-2
d
-1

547
+ 538
- 119
J bSi*Ge/Si opal
Ge non-opal burial flux,
pmol m
-2
d
-1

10
+ 278
- 136
J non-opalGe = J bGe – (J bSi*Ge/Si opal)
Ge non-opal fraction of
total burial
2 %
+ 20 %
- 42 %
J non-opal Ge / J bGe
Ge non-opal burial
fraction of opal rain
0.2 %
+ 5.5 %
- 3.2 %
J non-opal Ge /(J b Ge+J eGe)
a
These model output values were obtained using the “Main input” values from Table 5 and unadjusted data from Table 4.
b
Obtained by varying input parameters within the range specified in Table 5.
c
Surface-formed bSi contribution to total buried bSi.
d
Seawater-opal fractionation factor, defined as the ratio Ge/Si opal / Ge/Si seawater.
e
See text above for how these values were calculated. Km Ge-bottom has lower uncertainty estimates due to the fact that Km Si used to calculate it did
not have associated uncertainty.

49

Table 7. Comparison of Si and Ge fluxes at different locations
Location
Refe-
rences
Benthic flux
Ge/Si
a,b
,
µmol/mol
Benthic Si
flux
a,c
, mmol
m
-2
d
-1

bSi rain flux
a,d
,
mmol m
-2
d
-1

bSi burial
flux
a,e
,
mmol m
-2
d
-1

Fraction of
bSi rain
buried
Ge nonopal
burial flux
f
,
pmol m
-2
d
-1

Ge burial
fraction as
non-opal
g

f-ratio
h
The Gulf of
Mexico

Sta. 1
i
1 1.24 ± 0.13 4.11 ± 1.37
4.55
(3.07 – 6.36)
0.44
(0.3 – 0.9)

10%
10
(-125 – 288)
2%
(-40 – 22%)
0 ± 0.10
Sta. M 1 1.04 ± 0.04 0.59 ± 0.01 0.67 0.08 12% 9 9.5% 0.05 ± 0.03
Sta. G 1 0.88 ± 0.11 0.18 ± 0.04 0.23 0.045 20% 48 49% 0.20 ± 0.10
CA Borderlands

Santa Monica
Basin
2, 3, 4,
8, 9, 10
0.3
(0.2 - 1.8)
1.10
(1.1 - 2.3)
1.12
(0.25 – 2.5)
0.020 ± 0.002 1.8 ± 0.2% 461 97%
0.58
(-1.50 - 0.72)
San Pedro Basin
2, 5, 6,
7
0.3
(0.15 - 0.78)
1.62
(0.6 - 2.9)
1.7
(0.6 – 5.1)
0.08 ± 0.03 4.7 ± 1.8% 680 92%
0.58
(-0.08 - 0.79)
San Nicolas Basin 5, 7 0.5 ± 0.3 0.7 ± 0.1 0.76 ± 0.12 0.06 ± 0.02 7.9 ± 2.6% 154 78%
0.31
(-0.11 - 0.72)
South Atlantic

21GGC 8 0.03 0.019 0.24 0.22 92% 13 8% 0.96
22GGC 8 0.03 0.063 0.15 0.08 57% 44 43% 0.95
10GC 8 0.08 0.074 0.57 0.49 87% 38 10% 0.71
17GGC 8 0.71 2.49 4.44 1.95 44% 18 1% 0.01
13MC 8 0.69 1.48 2.93 1.45 50% 68 6% 0.06
References: (1) This study; (2) (McManus et al., 2003); (3) J. Baronas, unpublished data; (4) (Landry et al., 1992); (5) (Murnane et al., 1989); (6)
(Collins et al., 2011); (7) (Berelson et al., 1987); (8) (King et al., 2000); (9) (Christensen et al., 1994); (10) Boucher, 1984.
a
For CA Borderlands, range of all previously measured values, including uncertainty, is given in brackets.

b
GOM values obtained from slopes in Fig. 6. Other values from references.
c
GOM values from Table 2. Calculated as sum of Si burial and benthic fluxes for San Pedro and Santa Monica basins; other values from
references.
d
Sediment trap data for San Pedro and San Nicolas basins; calculated as the sum of benthic and burial fluxes for the other sites.
e
GOM values calculated from sediment accumulation rate and %bSi (see Table 3). %bSi not measured at Sta. M but assumed 1 ± 0.5% based on
the other two stations. Other values from references.
f
Model-calculated for Sta. 1 (Section 5.2). Calculated assuming Ge/Si opal = 0.72 for the South Atlantic and CA Borderlands sites, and Ge/Si opal =
1.1 µmol/mol for GOM Sta. M and Sta. G.
g
Defined as (non-opal Ge burial flux)/(total Ge burial flux).
h
As previously defined in Ref. 8, the fraction of the “potential” benthic Ge flux that is sequestered by the non-opal phase, or (Ge/Si opal –
Ge/Si benthic-flux) / (Ge/Si opal).
i
Values in brackets are model input and output ranges (Tables 5 and 6).

50

Figure 1. Map of sampling stations of cruise EN-497 that are discussed in this study. Bathymetry lines of
50 m and 500 m are given as operational definitions of continental shelf and continental slope,
respectively. The inset shows the location of the study area on a continental scale. Station details are
given in Table 1. Surface water salinity and bottom water O
2
concentration grids are taken from (Rabalais
and Turner, 2011). The hypoxic zone (< 2 mg/L = 60 µM O
2
) is denoted by the gray line.

51

Figure 2. Water column profiles of Ge and Si concentrations, as well as Ge/Si ratios, over the continental
shelf (Sta. 1 and 2), continental slope (Sta. 5), and offshore (Sta. G). Where not visible, the error bars are
smaller than the symbols.

52

Figure 3. Ge and Si relationship in the water column of the Northern Gulf of Mexico: a) Ge/Si ratio
plotted as a function of Si concentration at all stations. Inset shows a magnified region of the same data;
b) Ge plotted as a function of Si, with linear regression slopes for offshore Sta. G (solid line) and all shelf
station samples (dashed line). Symbols same as in panel (a). Slope uncertainties are reported as ± 2σ S.D.
Value in brackets was not included in slope calculation. Where not visible, the error bars are smaller than
the symbols. Where not visible, the error bars are smaller than the symbols.

53

Figure 4. Pore water profiles of dissolved Si, Ge, Fe
2+
, Mn
2+
, and NH
4
+
at Sta. 1 (Type A: cont. shelf
close to river plumes), Sta. 2 (Type B: cont. shelf far from river plumes), Sta. M (Type C: cont. slope),
and Sta. G (Type D: cont. rise). Pore water Ge/Si ratio profiles are plotted in the middle panel. Not the
scale changes for Si and Ge between different stations. Horizontal dashed lines mark the sediment-water
interface, and the sample above the line is the overlying bottom water. Vertical error bars represent the
range of sampling depth for composite aliquots. Horizontal error bars represent analytical uncertainty
(smaller than symbols where not visible). Fe
2+
, Mn
2+
, and NH
4
+
data for Sta. 1 and Sta. 2 were obtained
from a different core (but same deployment) than Si and Ge data.

54

Figure 5. Pore water profiles of dissolved Si, Fe
2+
, Mn
2+
, and NH
4
+
at several stations. Stations are
grouped according to different pore water profile types (see text above).

55

Figure 6. Core incubation experiment data. Right panels: Ge and Si concentration increase over time in
each core at a given station. Left panels: Ge concentration plotted as a function of Si concentration. Each
core is fit with a linear slope. The Ge/Si values given are the mean ± 1 S.D. of all the slopes at each given
station.

56

Figure 7. Si depletion and enrichment over the shelf: A) Silicic acid concentration plotted as a function of
salinity. Solid line is produced by mixing of Mississippi river water (Sal. = 0 psu, Si = 169 µM, out of
range) with GOM offshore water (Sta. G, 60 m depth) and is described by the equation. Samples falling
below the line indicate Si drawdown due to bSi production; samples falling above the line indicate Si
enrichment from dissolving bSi. B) Ratio of measured Si concentration to that predicted from the Si
cons

equation in panel A, plotted as depth profiles. Symbols same as in A. Inset shows surface water values in
more detail.

57

Figure 8. Box model of Sta. 1 water column and sediment Si and Ge cycling. Model output parameters
are given in bold, the rest are input parameters. “C” (concentration) and “J” (particulate flux) parameters
have both Ge and Si values (see Tables 5 and 6).

58

Figure 9. Pore water Ge/Si plotted as a function of 1/Si. Lines show conservative mixing slopes that the
pore water samples would fall on, if they were a result of mixing between overlying water and dissolving
bSi (see Table 3 for Ge/Si values; assumed Sta. 2 has Ge/Si
opal
= 1.2 µmol/mol identical to Sta. 1; 1/Si = 0
µM
-1

by definition) and overlying bottom water (Ge/Si and Si measured at each station; data from 2 m
above sediment-water interface used at Sta. 2; Si concentration defines the upper limit of the 1/Si axis).

59

8. REFERENCES
Aller, R.C., 2004, Conceptual models of early diagenetic processes: The muddy seafloor as an unsteady,
batch reactor: Journal of Marine Research, v. 62, no. 6, p. 815–835, doi:
10.1357/0022240042880837.
Aller, R.C., 1980, Quantifying solute distributions in the bioturbated zone of marine sediments by
defining an average microenvironment: Geochimica et Cosmochimica Acta, v. 44, no. 12, p. 1955–
1965, doi: 10.1016/0016-7037(80)90195-7.
Aller, R.C., 2014, Sedimentary Diagenesis, Depositional Environments, and Benthic Fluxes, in Holland,
H.D. and Turekian, K.K. eds., Treatise on Geochemistry, Elsevier, p. 293–334.
Aller, R., Blair, N., Xia, Q., and Rude, P., 1996, Remineralization rates, recycling, and storage of carbon
in Amazon shelf sediments: Continental Shelf Research, v. 16, no. 5/6, p. 753–786.
Aller, R.C., Hannides, A., Heilbrun, C., and Panzeca., C., 2004a, Coupling of early diagenetic processes
and sedimentary dynamics in tropical shelf environments: the Gulf of Papua deltaic complex:
Continental Shelf Research, v. 24, no. 19, p. 2455–2486, doi: 10.1016/j.csr.2004.07.018.
Aller, R.C., Heilbrun, C., Panzeca, C., Zhu, Z., and Baltzer, F., 2004b, Coupling between sedimentary
dynamics, early diagenetic processes, and biogeochemical cycling in the Amazon–Guianas mobile
mud belt: coastal French Guiana: Marine Geology, v. 208, no. 2-4, p. 331–360, doi:
10.1016/j.margeo.2004.04.027.
Allison, M.A., Kineke, G.C., Gordon, E.S., and Goñi, M.A., 2000, Development and reworking of a
seasonal flood deposit on the inner continental shelf off the Atchafalaya River: v. 20, p. 2267–2294.
Anders, A.M., Sletten, R., Derry, L., and Hallet, B., 2003, Germanium/silicon ratios in the Copper River
Basin, Alaska: Weathering and partitioning in periglacial versus glacial environments: Journal of
Geophysical Research, v. 108, no. F1, p. 6005, doi: 10.1029/2003JF000026.
Andersen, M.B., Vance, D., Archer, C., Anderson, R.F., Ellwood, M.J., and Allen, C.S., 2011, The Zn
abundance and isotopic composition of diatom frustules, a proxy for Zn availability in ocean surface
seawater: Earth and Planetary Science Letters, v. 301, no. 1-2, p. 137–145, doi:
10.1016/j.epsl.2010.10.032.
Bareille, G., Labracherie, M., Mortlock, R.A., Maier-Reimer, E., and Froelich, P.N., 1998, A test of
(Ge/Si) opal as a paleorecorder of (Ge/Si) seawater: Geology, v. 26, no. 2, p. 179–182.
Berelson, W., and Hammond, D., 1986, The calibration of a new free-vehicle benthic flux chamber for
use in the deep sea: Deep Sea Research Part A. Oceanographic Research Papers, v. 33, no. 10, p.
1439–1454.
Berelson, W.., Hammond, D.., and Johnson, K.., 1987, Benthic fluxes and the cycling of biogenic silica
and carbon in two southern California borderland basins: Geochimica et Cosmochimica Acta, v. 51,
p. 1345–1363, doi: 10.1016/0016-7037(87)90320-6.

60

Bernstein, L., 1985, Germanium geochemistry and mineralogy: Geochimica et Cosmochimica Acta, v.
49, p. 2409–2422.
Bernstein, R., and Waychunas, G.A., 1987, Germanium crystal chemistry in hematite and goethite from
the Apex Mine , Utah , and some new data on germanium in aqueous solution and in stottite: v. 51,
no. 1959.
Bianchi, T.S., DiMarco, S.F., Cowan, J.H., Hetland, R.D., Chapman, P., Day, J.W., and Allison, M. a,
2010, The science of hypoxia in the Northern Gulf of Mexico: a review.: The Science of the Total
Environment, v. 408, p. 1471–1484, doi: 10.1016/j.scitotenv.2009.11.047.
Blecker, S.W., King, S.L., Derry, L. a., Chadwick, O. a., Ippolito, J. a., and Kelly, E.F., 2007, The ratio of
germanium to silicon in plant phytoliths: quantification of biological discrimination under controlled
experimental conditions: Biogeochemistry, v. 86, no. 2, p. 189–199, doi: 10.1007/s10533-007-9154-
7.
Bower, C., and Holm-Hansen, T., 1980, A salicylate-hypochlorite method for determining ammonia in
seawater: Canadian Journal of Fisheries and Aquatic Science, v. 37, p. 794–798.
Burdige, D.J., 1993, The biogeochemistry of manganese and iron reduction in marine sediments: Earth-
Science Reviews, v. 35, no. 3, p. 249–284, doi: 10.1016/0012-8252(93)90040-E.
Chillrud, S.N., Pedrozo, F., Temporetti, P., Planas, H., and Froelich, P., 1994, Chemical weathering of
phosphate and germanium in glacial meltwater streams: Effects of subglacial pyrite oxidation:
Limnology and Oceanography, v. 39, no. 5, p. 1130–1140, doi: 10.4319/lo.1994.39.5.1130.
Chong, L.S., Berelson, W.M., McManus, J., Hammond, D.E., Rollins, N.E., and Yager, P.L., 2014,
Carbon and biogenic silica export influenced by the Amazon River Plume: Patterns of
remineralization in deep-sea sediments: Deep Sea Research Part I: Oceanographic Research Papers,
v. 85, p. 124–137, doi: 10.1016/j.dsr.2013.12.007.
Christensen, C., Gorsline, D., Hammond, D., and Lund, S., 1994, Non-annual laminations and expansion
of anoxic basin-floor conditions in Santa Monica Basin, California Borderland, over the past four
centuries: Marine Geology, v. 116, p. 399–418.
Collins, L.E., Berelson, W., Hammond, D.E., Knapp, A., Schwartz, R., and Capone, D., 2011, Particle
fluxes in San Pedro Basin, California: A four-year record of sedimentation and physical forcing:
Deep Sea Research I, v. 58, no. 8, p. 898–914, doi: 10.1016/j.dsr.2011.06.008.
Corbett, D., Dail, M., and McKee, B., 2007, High-frequency time-series of the dynamic sedimentation
processes on the western shelf of the Mississippi River Delta: Continental Shelf Research, v. 27, p.
1600–1615, doi: 10.1016/j.csr.2007.01.025.
Corbett, D., McKee, B., and Allison, M., 2006, Nature of decadal-scale sediment accumulation on the
western shelf of the Mississippi River delta: Continental Shelf Research, v. 26, p. 2125–2140, doi:
10.1016/j.csr.2006.07.012.
Cornelis, J.-T., Delvaux, B., Cardinal, D., André, L., Ranger, J., and Opfergelt, S., 2010, Tracing
mechanisms controlling the release of dissolved silicon in forest soil solutions using Si isotopes and

61

Ge/Si ratios: Geochimica et Cosmochimica Acta, v. 74, no. 14, p. 3913–3924, doi:
10.1016/j.gca.2010.04.056.
Cornelis, J.-T., Weis, D., Lavkulich, L., Vermeire, M.-L., Delvaux, B., and Barling, J., 2014, Silicon
isotopes record dissolution and re-precipitation of pedogenic clay minerals in a podzolic soil
chronosequence: Geoderma, v. 235-236, p. 19–29, doi: 10.1016/j.geoderma.2014.06.023.
Delvigne, C., Opfergelt, S., Cardinal, D., Delvaux, B., and André, L., 2009, Distinct silicon and
germanium pathways in the soil-plant system: Evidence from banana and horsetail: Journal of
Geophysical Research, v. 114, no. G2, p. G02013, doi: 10.1029/2008JG000899.
DeMaster, D.J., 1981, The supply and accumulation of silica in the marine environment: Geochimica et
Cosmochimica Acta, v. 45, p. 1715–1732, doi: 10.1016/0016-7037(81)90006-5.
DeMaster, D., Knapp, G., and Nittrouer, C., 1983, Biological uptake and accumulation of silica on the
Amazon continental shelf: Geochimica et Cosmochimica Acta, v. 47, p. 1713–1723.
Derry, L. a, Kurtz, A.C., Ziegler, K., and Chadwick, O. a, 2005, Biological control of terrestrial silica
cycling and export fluxes to watersheds.: Nature, v. 433, p. 728–31, doi: 10.1038/nature03299.
Derry, L. a., Pett-Ridge, J.C., Kurtz, A.C., and Troester, J.W., 2006, Ge/Si and 87Sr/86Sr tracers of
weathering reactions and hydrologic pathways in a tropical granitoid system: Journal of
Geochemical Exploration, v. 88, no. 1-3, p. 271–274, doi: 10.1016/j.gexplo.2005.08.054.
Dixit, S., Van Cappellen, P., and van Bennekom, A.J., 2001, Processes controlling solubility of biogenic
silica and pore water build-up of silicic acid in marine sediments: Marine Chemistry, v. 73, no. 3-4,
p. 333–352, doi: 10.1016/S0304-4203(00)00118-3.
Ellwood, M.J., Kelly, M., Maher, W. a., and De Deckker, P., 2006, Germanium incorporation into sponge
spicules: Development of a proxy for reconstructing inorganic germanium and silicon
concentrations in seawater: Earth and Planetary Science Letters, v. 243, p. 749–759, doi:
10.1016/j.epsl.2006.01.016.
Ellwood, M.J., and Maher, W. a, 2003, Germanium cycling in the waters across a frontal zone: the
Chatham Rise, New Zealand: Marine Chemistry, v. 80, no. 2-3, p. 145–159, doi: 10.1016/S0304-
4203(02)00115-9.
Esther, T. a., Hammond, D.E., Hautala, S.L., Paul Johnson, H., Schwartz, R.J., and Paukert, A.N., 2010,
Evaluation of the budget for silicic acid in Cascadia Basin deep water: Deep Sea Research Part I:
Oceanographic Research Papers, v. 57, no. 5, p. 677–686, doi: 10.1016/j.dsr.2010.02.002.
Froelich, P., Blanc, V., Mortlock, R., Chillrud, S., Dunstan, W., Udomkit, A., and Peng, T., 1992, River
fluxes of dissolved silica to the ocean were higher during glacials: Ge/Si in diatoms, rivers, and
oceans: Paleoceanography, v. 7, no. 6, p. 739–767.
Froelich, P., Hambrick, G., Andreae, M., Mortlock, R., and Edmond, J., 1985a, The geochemistry of
inorganic germanium in natural waters: Journal of Geophysical Research, v. 90, no. C1, p. 1133–
1141.

62

Froelich, P.N., Hambrick, G.A., Kaul, L.W., Byrd, J.T., and Lecointe, O., 1984, Geochemical behavior of
inorganic geranium in an unperturbed estuary: Geochimica et Cosmochimica Acta, v. 49, p. 519–
524.
Froelich, P.N., Kaul, L.W., Byrd, J.T., Andreae, M.O., and Roe, K.K., 1985b, Arsenic, barium,
germanium, tin, dimethylsulfide and nutrient biogeochemistry in Charlotte Harbor, Florida, a
phosphorus-enriched estuary: Estuarine, Coastal and Shelf Science, v. 20, no. 3, p. 239–264, doi:
10.1016/0272-7714(85)90041-1.
Froelich, P., Mortlock, R., and Shemesh, A., 1989, Inorganic germanium and silica in the Indian Ocean:
biological fractionation during (Ge/Si) Opal formation: Global biogeochemical cycles, v. 3, no. 1, p.
79–88.
Gallinari, M., Ragueneau, O., and Corrin, L., 2002, The importance of water column processes on the
dissolution properties of biogenic silica in deep-sea sediments I. Solubility: Geochimica et
Cosmochimica Acta, v. 66, no. 15, p. 2701–2717.
Gordon, E.S., and Goñi, M. a., 2004, Controls on the distribution and accumulation of terrigenous organic
matter in sediments from the Mississippi and Atchafalaya river margin: Marine Chemistry, v. 92,
no. 1-4, p. 331–352, doi: 10.1016/j.marchem.2004.06.035.
Hammond, D.E., Cummins, K.M., Mcmanus, J., Berelson, W.M., and Smith, G., 2004a, Methods for
measuring benthic nutrient flux on the California Margin : Comparing shipboard core incubations to
in situ lander results: Limnology and Oceanography: Methods, v. 2, p. 146–159.
Hammond, D.E., McManus, J., and Berelson, W., 2004b, Oceanic germanium/silicon ratios: Evaluation
of the potential overprint of temperature on weathering signals: Paleoceanography, v. 19, p.
PA2016, doi: 10.1029/2003PA000940.
Hammond, D.E., Mcmanus, J., Berelson, W.M., and Kilgore, T.E., 1996, Early diagenesis of organic
material in equatorial Pacific sediments : stoichiometry and kinetics: Deep Sea Research II, v. 43,
no. 4-6, p. 1365–1412.
Hammond, D., McManus, J., Berelson, W., Meredith, C., Klinkhammer, P., and Coale, K., 2000,
Diagenetic fractionation of Ge and Si in reducing sediments: the missing Ge sink and a possible
mechanism to cause glacial/interglacial variations in oceanic Ge/Si: Geochimica et Cosmochimica
Acta, v. 64, no. 14, p. 2453–2465.
Hendry, K.R., and Brzezinski, M. a., 2014, Using silicon isotopes to understand the role of the Southern
Ocean in modern and ancient biogeochemistry and climate: Quaternary Science Reviews, v. 89, p.
13–26, doi: 10.1016/j.quascirev.2014.01.019.
Holmes, C., 1976, Distribution regional variation, and geochemical coherence of selected elements in the
sediments of the Central Gulf of Mexico: Geological Survey professional paper,.
Jendrzejewski, J.P., and Hart, G.F., 1978, Distribution of Siliceous Microfossils in Surficial Bottom
Sediments of the Gulf of Mexico: Palynology, v. 2, p. 159–166.

63

Kamatani, A., and Oku, O., 2000, Measuring biogenic silica in marine sediments: Marine Chemistry, v.
68, p. 219–229, doi: 10.1016/S0304-4203(99)00079-1.
Khalil, K., Rabouille, C., Gallinari, M., Soetaert, K., DeMaster, D.J., and Ragueneau, O., 2007,
Constraining biogenic silica dissolution in marine sediments: A comparison between diagenetic
models and experimental dissolution rates: Marine Chemistry, v. 106, no. 1-2, p. 223–238, doi:
10.1016/j.marchem.2006.12.004.
King, S., Froelich, P., and Jahnke, R., 2000, Early diagenesis of germanium in sediments of the Antarctic
South Atlantic: in search of the missing Ge sink: Geochimica et Cosmochimica Acta, v. 64, no. 8, p.
1375–1390.
Kolodny, Y., and Halicz, L., 1988, The geochemistry of germanium in deep-sea cherts: Geochimica et
Cosmochimica Acta, v. 52, p. 2333–2336.
Koning, E., Brummer, G.-J., van Raaphorst, W., van Bennekom, J., Helder, W., and van Iperen, J., 1997,
Settling, dissolution and burial of biogenic silica in the sediments off Somalia (northwestern Indian
Ocean): Deep Sea Research II, v. 44, no. 6-7, p. 1341–1360.
Kurtz, A., Derry, L., and Chadwick, O., 2002, Germanium-silicon fractionation in the weathering
environment: Geochimica et cosmochimica acta, v. 66, no. 9, p. 1525–1537.
Landry, M.R., Peterson, W.K., and Andrews, C.C., 1992, Particulate flux in the water column overlying
Santa Monica Basin: Progress in Oceanography, v. 30, no. 1-4, p. 167–195, doi: 10.1016/0079-
6611(92)90012-O.
Lin, H.-L., and Chen, C.-J., 2002, A late Pliocene diatom Ge/Si record from the Southeast Atlantic:
Marine Geology, v. 180, p. 151–161, doi: 10.1016/S0025-3227(01)00211-0.
Lin, S., and Morse, J., 1991, Sulfate reduction and iron sulfide mineral formation in Gulf of Mexico
anoxic sediments: American Journal of Science, v. 291, p. 55–89.
Loucaides, S., Cappellen, P., Roubeix, V., Moriceau, B., and Ragueneau, O., 2011, Controls on the
Recycling and Preservation of Biogenic Silica from Biomineralization to Burial: Silicon, v. 4, p. 7–
22, doi: 10.1007/s12633-011-9092-9.
Loucaides, S., Michalopoulos, P., Presti, M., Koning, E., Behrends, T., and Van Cappellen, P., 2010,
Seawater-mediated interactions between diatomaceous silica and terrigenous sediments: Results
from long-term incubation experiments: Chemical Geology, v. 270, p. 68–79, doi:
10.1016/j.chemgeo.2009.11.006.
Lugolobi, F., Kurtz, A.C., and Derry, L. a., 2010, Germanium–silicon fractionation in a tropical, granitic
weathering environment: Geochimica et Cosmochimica Acta, v. 74, no. 4, p. 1294–1308, doi:
10.1016/j.gca.2009.11.027.
McManus, J., Hammond, D.E., Berelson, W.M., Kilgore, T.E., Demaster, D.J., Ragueneau, O.G., and
Collier, R.W., 1995, Early diagenesis of biogenic opal: Dissolution rates, kinetics, and
paleoceanographic implications: Deep Sea Research II, v. 42, no. 2-3, p. 871–903.

64

McManus, J., Hammond, D.E., Cummins, K., Klinkhammer, G.P., and Berelson, W.M., 2003, Diagenetic
Ge-Si fractionation in continental margin environments: further evidence for a nonopal Ge sink:
Geochimica et Cosmochimica Acta, v. 67, no. 23, p. 4545–4557, doi: 10.1016/S0016-
7037(03)00385-5.
Michalopoulos, P., and Aller, R.C., 2004, Early diagenesis of biogenic silica in the Amazon delta:
alteration, authigenic clay formation, and storage: Geochimica et Cosmochimica Acta, v. 68, no. 5,
p. 1061–1085, doi: 10.1016/j.gca.2003.07.018.
Michalopoulos, P., and Aller, R.C., 1995, Rapid Clay Mineral Formation in Amazon Delta Sediments :
Reverse Weathering and Oceanic Elemental Cycles: Science, v. 270, no. 5236, p. 614–617.
Michalopoulos, P., Aller, R., and Reeder, R., 2000, Conversion of diatoms to clays during early
diagenesis in tropical, continental shelf muds: Geology, v. 28, p. 1095–1098, doi: 10.1130/0091-
7613(2000)28<1095.
Morley, D.W., Leng, M.J., Mackay, A.W., Sloane, H.J., Rioual, P., and Battarbee, R.W., 2004, Cleaning
of lake sediment samples for diatom oxygen isotope analysis: Journal of Paleolimnology, v. 31, no.
3, p. 391–401, doi: 10.1023/B:JOPL.0000021854.70714.6b.
Mortlock, R., Charles, C., Froelich, P., Zibello, M., Saltzman, J., Hays, J., and Burckle, L., 1991,
Evidence for lower productivity in the Antarctic Ocean during the last glaciation: Nature, v. 351, p.
220–223.
Mortlock, R., and Froelich, P., 1989, A simple method for the rapid determination of biogenic opal in
pelagic marine sediments: Deep Sea Research, v. 36, no. 9, p. 1415–1426.
Mortlock, R., and Froelich, P., 1987, Continental weathering of germanium: Ge/Si in the global river
discharge: Geochimica et Cosmochimica Acta, v. 51, p. 2075–2082.
Mortlock, R., and Froelich, P., 1996, Determination of germanium by isotope dilution-hydride generation
inductively coupled plasma mass spectrometry: Analytica chimica acta, v. 332, p. 277–284.
Mortlock, R., and Froelich, P., 1986, Hydrothermal germanium over the southern East Pacific Rise:
Science, v. 231, no. 4733, p. 43–45.
Mortlock, R.A., Froelich, P.N., Feely, R.A., Massoth, G.J., Butterfield, D.A., and Lupton, J.E., 1993,
Silica and germanium in Pacific Ocean hydrothermal vents and plumes: Earth and Planetary Science
Letters, v. 119, no. 3, p. 365–378, doi: 10.1016/0012-821X(93)90144-X.
Mullin, J., and Riley, J., 1955, The colorimetric determination of silicate with special reference to sea and
natural waters: Analytica Chimica Acta, v. 12, p. 162–175.
Murnane, R.J., Leslie, B., Hammond, D.E., and Stallard, R.F., 1989, Germanium geochemistry in the
Southern California Borderlands: Geochimica et Cosmochimica Acta, v. 53, p. 2873–2882.
Murnane, R., and Stallard, R., 1990, Germanium and silicon in rivers of the Orinoco drainage basin:
Nature, v. 344, p. 749–752.

65

Murnane, R.J., and Stallard, R.F., 1988, Germanium/silicon fractionation during biogenic opal formation:
Paleoceanography, v. 3, no. 4, p. 461–469.
Nelson, D.M., and Dortch, Q., 1996, Silicic acid depletion and silicon limitation in the plume of the
Mississippi River : evidence from kinetic studies in spring and summer: Marine Ecology Progress
Series, v. 136, p. 163–178.
Opfergelt, S., Cardinal, D., André, L., Delvigne, C., Bremond, L., and Delvaux, B., 2010, Variations of
δ30Si and Ge/Si with weathering and biogenic input in tropical basaltic ash soils under monoculture:
Geochimica et Cosmochimica Acta, v. 74, no. 1, p. 225–240, doi: 10.1016/j.gca.2009.09.025.
Pakulski, J.D., Benner, R., Whitledge, T., Amon, R., Eadie, B., Cifuentes, L., Ammerman, J., and
Stockwell, D., 2000, Microbial Metabolism and Nutrient Cycling in the Mississippi and Atchafalaya
River Plumes: Estuarine, Coastal and Shelf Science, v. 50, p. 173–184, doi:
10.1006/ecss.1999.0561.
Pokrovsky, O.S., Pokrovski, G.S., Schott, J., and Galy, a., 2006, Experimental study of germanium
adsorption on goethite and germanium coprecipitation with iron hydroxide: X-ray absorption fine
structure and macroscopic characterization: Geochimica et Cosmochimica Acta, v. 70, no. 13, p.
3325–3341, doi: 10.1016/j.gca.2006.04.012.
Presti, M., and Michalopoulos, P., 2008, Estimating the contribution of the authigenic mineral component
to the long-term reactive silica accumulation on the western shelf of the Mississippi River Delta:
Continental Shelf Research, v. 28, p. 823–838, doi: 10.1016/j.csr.2007.12.015.
Rabalais, N., and Turner, R., 2011, Press release. Louisiana Universities Marine Consortium.
http://www.gulfhypoxia.net.
Rabalais, N., Turner, R., Dortch, Q., Justic, D., Bierman, V.J., and Wiseman, W.J., 2002, Nutrient-
enhanced productivity in the northern Gulf of Mexico: past, present and future: Hydrobiologia, v.
475/476, p. 39–63.
Rabalais, N., Turner, R., and Justic, D., 1996, Nutrient changes in the Mississippi River and system
responses on the adjacent continental shelf: Estuaries, v. 19, no. 2, p. 386–407.
Schink, D., Guinasso Jr., N., and Fanning, K., 1975, Processes affecting the concentration of silica at the
sediment‐water interface of the Atlantic Ocean: Journal of Geophysical Research, v. 80, no. 21, p.
3013–3031.
Shemesh, A., Mortlock, R.A., and Froelich, P.N., 1989, Late cenozoic Ge/Si record of marine biogenic
opal: Implications for variations of riverine fluxes to the ocean: Paleoceanography, v. 4, no. 3, p.
221–234.
Shemesh, A., Mortlock, R., Smith, R., and Froelich, P., 1988, Determination of Ge/Si in marine siliceous
microfossils: Separation, cleaning and dissolution of diatoms and radiolaria: Marine chemistry, v.
25, p. 305–323.
Sparks, J.P., Chandra, S., Derry, L. a., Parthasarathy, M. V., Daugherty, C.S., and Griffin, R., 2010,
Subcellular localization of silicon and germanium in grass root and leaf tissues by SIMS: evidence

66

for differential and active transport: Biogeochemistry, v. 104, no. 1-3, p. 237–249, doi:
10.1007/s10533-010-9498-2.
Sutton, J.N., 2011, Germanium/Silicon and Silicon Isotope Fractionation by Marine Diatoms and Sponges
and Utility as Tracers of Silicic Acid Utilization.
Sutton, J., Ellwood, M.J., Maher, W. a., and Croot, P.L., 2010, Oceanic distribution of inorganic
germanium relative to silicon: Germanium discrimination by diatoms: Global Biogeochemical
Cycles, v. 24, p. GB2017, doi: 10.1029/2009GB003689.
Thamatrakoln, K., and Hildebrand, M., 2008, Silicon uptake in diatoms revisited: a model for saturable
and nonsaturable uptake kinetics and the role of silicon transporters.: Plant physiology, v. 146, no. 3,
p. 1397–407, doi: 10.1104/pp.107.107094.
Tosca, R., Tosca, N., and Galy, A., 2014, Germanium incorporation into authigenic marine silicate
minerals : an experimental approach: Mineralogical Magazine, v. 77, no. 5.
Trefry, J.H., and Presley, B.J., 1982, Manganese fluxes from Mississippi Delta sediments: Geochimica et
Cosmochimica Acta, v. 46, no. 10, p. 1715–1726, doi: 10.1016/0016-7037(82)90112-0.
Tréguer, P.J., and De La Rocha, C.L., 2013, The World Ocean Silica Cycle: Annual review of marine
science, v. 5, p. 477–501, doi: 10.1146/annurev-marine-121211-172346.
Tribovillard, N., Bout-Roumazeilles, V., Riboulleau, a., Baudin, F., Danelian, T., and Riquier, L., 2011,
Transfer of germanium to marine sediments: Insights from its accumulation in radiolarites and
authigenic capture under reducing conditions. Some examples through geological ages: Chemical
Geology, v. 282, no. 3-4, p. 120–130, doi: 10.1016/j.chemgeo.2011.01.015.
Turner, R.E., Milan, C.S., and Rabalais, N.N., 2004, A retrospective analysis of trace metals, C, N and
diatom remnants in sediments from the Mississippi River delta shelf.: Marine pollution bulletin, v.
49, p. 548–556, doi: 10.1016/j.marpolbul.2004.03.013.
Turner, R., and Rabalais, N., 1994, Coastal eutrophication near the Mississippi river delta: Nature, v. 368,
p. 619–621.
Turner, R., Rabalais, N., Alexander, R., McIsaac, G., and Howarth, R., 2007, Characterization of nutrient,
organic carbon, and sediment loads and concentrations from the Mississippi River into the Northern
Gulf of Mexico: Estuaries and Coasts, v. 30, no. 5, p. 773–790.
Walker, N.D., Wiseman, W.J., Rouse, L.J., and Babin, A., 2005, Effects of River Discharge, Wind Stress,
and Slope Eddies on Circulation and the Satellite-Observed Structure of the Mississippi River
Plume: Journal of Coastal Research, v. 21, no. 6, p. 1228–1244, doi: 10.2112/04-0347.1.
Wheat, C.G., and McManus, J., 2008, Germanium in mid-ocean ridge flank hydrothermal fluids:
Geochemistry, Geophysics, Geosystems, v. 9, no. 3, p. 1–16, doi: 10.1029/2007GC001892.
Wheat, C.G., and McManus, J., 2005, The potential role of ridge-flank hydrothermal systems on oceanic
germanium and silicon balances: Geochimica et Cosmochimica Acta, v. 69, no. 8, p. 2021–2029,
doi: 10.1016/j.gca.2004.05.046.

67

White, A.F., Vivit, D. V., Schulz, M.S., Bullen, T.D., Evett, R.R., and Aagarwal, J., 2012, Biogenic and
pedogenic controls on Si distributions and cycling in grasslands of the Santa Cruz soil
chronosequence, California: Geochimica et Cosmochimica Acta, v. 94, p. 72–94, doi:
10.1016/j.gca.2012.06.009.
Wright, L., and Coleman, J., 1971, Effluent expansion and interfacial mixing in the presence of a salt
wedge, Mississippi River delta: Journal of Geophysical Research, v. 76, no. 36, p. 8649–8661, doi:
10.1029/JC076i036p08649.

68

APPENDIX A: Additional data and figures
Table A1. Summary of water column Ge and Si concentrations
Station
Bottom
depth, m
Sample
name
Sample
depth, m
Ge, pM ± Si, µM ±
Ge/Si,
µmol/mol
±
Sta.
MR
Miss_Riv

264.0 7.9 169.3 5.1 1.56 0.07
Sta. 1 30 CTD-5-12 0.7 15.5 0.5 3.2 0.4 4.93 0.64

CTD-5-11 4.0 9.3 0.5 3.2 0.4 2.86 0.38

CTD-5-10 6.1 19.6 0.6 7.7 0.4 2.55 0.15

CTD-5-9 10.0 28.9 0.9 14.1 0.4 2.05 0.09

CTD-5-8 12.0 42.7 1.3 20.2 0.6 2.11 0.09

CTD-5-6 16.0 83.7 2.5 37.3 1.1 2.24 0.10

CTD-5-5 18.0 80.4 2.4 37.7 1.1 2.14 0.09

CTD-5-4 20.3 75.6 2.3 36.8 1.1 2.06 0.09

CTD-5-3 22.0 72.9 2.2 35.8 1.1 2.04 0.09

CTD-5-2 24.0 70.0 2.1 30.8 0.9 2.27 0.10

CTD-5-1 24.4 66.6 2.0 30.9 0.9 2.16 0.09
Sta. 1 38 CTD-6 (30m) 30.0 45.7 1.4 25.1 0.8 1.82 0.08
Sta. 2 24 CTD-10-11 0.3 29.6 0.9 2.1 0.4 13.9 2.6

CTD-10-10 2.1

2.2 0.4

CTD-10-9 6.1

1.9 0.4

CTD-10-8 8.1

2.2 0.4

CTD-10-7 10.3 46.3 1.4 1.8 0.4 25.1 5.5

CTD-10-6 12.3

2.8 0.4

CTD-10-5 14.3

4.4 0.4

CTD-10-4 16.1

10.7 0.4

CTD-10-3 18.2 60.5 1.8 18.4 0.6 3.29 0.14

CTD-10-2 18.9

20.5 0.6

CTD-10-1 20.7

31.2 0.9

Sta. 5 89 CTD-23-12

1.4 0.4

CTD-23-11 9.3 2.4 0.5 1.2 0.4 2.05 0.83

CTD-23-10 19.3

1.2 0.4

CTD-23-9 29.2

1.0 0.4

CTD-23-8 39.0

1.5 0.4

CTD-23-7 49.1 2.7 0.5 1.3 0.4 2.12 0.76

CTD-23-6 59.2

2.1 0.4

CTD-23-5 62.3

1.9 0.4

CTD-23-4 69.2

2.9 0.4

CTD-23-3 75.3 9.3 0.5 5.0 0.4 1.87 0.18

CTD-23-2 83.2

6.3 0.4

CTD-23-1 89.3

6.9 0.4

Sta. 9 22 CTD-45-7 2.3 42.7 1.3 10.0 0.4 4.28 0.21

69

CTD-45-8 3.0 21.1 0.6 9.5 0.4 2.23 0.12

CTD-45-1 17.6

33.7 1.0

Sta. G 2153 CTD-32-10 219.6 2.1 0.5 9.1 0.4 0.23 0.06

CTD-32-9 259.1

10.6 0.4

CTD-32-8 279.7

11.4 0.4

CTD-32-7 403.0 12.0 0.5 16.6 0.5 0.72 0.04

CTD-32-6 1199.5

25.8 0.8

CTD-32-5 2120.6 21.8 0.7 25.3 0.8 0.86 0.04

CTD-32-2 2120.6 20.9 0.6 25.3 0.8 0.82 0.03

CTD-32-3 2120.6 21.9 0.7 25.5 0.8 0.86 0.04
Sta. G 2150 CTD-33-12 29.2

0.3 0.4

CTD-33-11 59.0 3.3 0.5 1.9 0.4 1.77 0.46

CTD-33-10 99.2 11.3 0.5 3.6 0.4 3.09 0.36

CTD-33-9 149.5

5.3 0.4

CTD-33-5 495.8 16.5 0.5 19.4 0.6 0.85 0.04

CTD-33-1 749.5 14.9 0.4 24.9 0.7 0.60 0.03

70

Table A2. Pore water concentrations
Sample name
Depth,
cm
Ge,
pM ±
Si,
µM ±
Ge/Si
a
,
µmol/mol ± Fe, µM ±
Mn,
µM ±
NH 4
+
,
µM
Sta. 1 (MC-2 core C,
composite) 0 51.4 1.5 22.9 0.7 2.25 0.10

1.5 204.0 6.1 214.3 6.4 0.95 0.04 25.5

60.6

49.1

4.5 741.4 22.2 174.4 5.2 4.25 0.18 37.6

31.3

63.1

8.0 429.0 12.9 133.8 4.0 3.21 0.14 11.7

24.2

46.9

12.5 207.9 6.2 132.7 4.0 1.57 0.07 1.9

19.1

39.7
Sta. 1 (MC-2*) 1

14.8 0.2 70.0 0.9

2

36.1 0.2 51.1 0.4

3

46.9 0.2 39.3 0.3

4

33.8 0.4 37.7 0.5

5

41.3 0.5 25.0 0.3

6

17.5 0.1 25.4 0.2

7

11.6 0.2 11.3 0.1

8

9.7 0.1 26.7 0.1

9

13.7 0.1 34.6 0.1

10

13.3 0.2 34.2 0.7

11

1.9 0.0 20.8 1.3

13

1.9 0.0 17.4 0.8

14

0.9 0.0 14.6 0.7

15

0.9 0.0 12.4 0.6

16

2.1 0.0 10.2 0.5

17

1.6 0.0 15.8 0.8
Sta. 2 (MC-6 cores
C-E, composite) 0-3 385.8 11.6 312.5 9.4 1.23 0.05 152.0 2.5

115.4

3-6 639.2 19.2 461.8 13.9 1.38 0.06 133.2 4.8

203.0

6-10 516.8 15.5 521.3 15.6 0.99 0.04 102.6 5.0

246.9

10-15 325.9 9.8 520.5 15.6 0.63 0.03 83.7 3.9

263.6

15-20 384.5 11.5 435.5 13.1 0.88 0.04 45.7 8.4

262.3
Sta. 2 (MC-5*) 0

31.4 0.9

0.4 0.0 0.0 0.0 0.1

1

304.6 9.1

126.4 1.4 171.1 1.4 85.6

2

150.4 4.5

132.6 0.6 132.8 0.7 51.3

3

456.7 13.7

98.5 0.4 143.9 1.4 164.5

4

401.8 12.1

108.0 1.6 143.4 1.7 171.6

5

486.3 14.6

92.1 0.2 122.9 0.9 198.0

6

502.7 15.1

81.6 0.5 116.1 0.3 202.9

7

84.8 1.0 121.7 1.7

8

579.2 17.4

74.8 0.7 103.8 1.0 251.2

9

63.4 1.1 82.2 1.3

10

583.5 17.5

73.5 1.3 111.3 1.8 259.4

12

50.3 0.4 83.1 0.1

14

529.9 15.9

46.9 0.3 84.2 0.8 243.2

71

16

469.0 14.1

43.2 0.5 85.1 1.0 222.8

18

353.4 10.6

19.8 0.8 6.2 0.6 191.6

20

20.2 0.5 2.1 0.1

23

204.3 6.1

9.9 0.3 0.5 0.1 148.6

26

160.2 4.8

7.6 0.3 0.6 0.1 147.6

29

160.2 4.8

5.8 0.2 0.3 0.0 155.5

32

6.9 0.0 0.5 0.1

33

164.4 4.9

4.0 0.0 0.2 0.0 149.1
Sta. M (MC-18 cores
D-H, composite) 0 20.8 0.6 21.5 0.6 0.96 0.04

0-3 73.2 2.2 131.2 3.9 0.56 0.02 0.1

127.8

3-6 69.8 2.1 132.2 4.0 0.53 0.02 0.2

235.5

6-13 51.3 1.5 103.9 3.1 0.49 0.02 31.0

254.6

15-27 54.0 1.6 128.2 3.8 0.42 0.02 36.2

133.4

26-39 80.8 2.4 168.2 5.0 0.48 0.02 59.4

112.3

0.113
Sta. M (MC-18 core
E) 0.5

137.9 4.1

0.1 0.0 58.5 1.3 1.6

1.5

141.8 4.3

0.1 0.0 104.9 1.9 4.5

2.5

132.5 4.0

0.1 0.0 220.1 4.9 15.3

3.5

135.3 4.1

0.2 0.0 190.4 0.5 12.3

4.5

135.7 4.1

0.2 0.0 242.3 6.2 24.6

5.5

123.5 3.7

0.2 0.0 273.7 2.0 25.4

7.5

105.4 3.2

15.4 0.2 270.2 3.7 46.1

9.5

118.4 3.6

26.1 0.4 261.3 4.2 61.5

11.5

130.3 3.9

51.6 1.3 232.2 5.4 73.8

16.5

78.9 2.4

10.6 0.1 134.4 1.6 105.0

22.5

154.7 4.6

65.0 1.0 150.1 3.0 124.9

26.5

117.5 3.5

33.1 0.6 115.6 1.5 145.9

30.5

152.4 4.6

51.9 1.3 114.3 1.8 159.7

33.5

170.0 5.1

66.9 0.7 110.4 0.3 167.1

34.5

188.7 5.7

65.3 1.4 110.1 1.5 171.1

38.5

220.8 6.6

39.4 1.1 108.7 2.3 192.6
Sta. G (MC-15 cores
D-H, composite) 0 26.9 0.8 28.9 0.9 0.93 0.04

0-3 56.5 1.7 89.7 2.7 0.63 0.03

3-6 62.0 1.9 127.6 3.8 0.49 0.02

6-13 57.2 1.7 151.7 4.5 0.38 0.02

14-27 53.0 1.6 166.5 5.0 0.32 0.01

26-39 50.7 1.5 163.5 4.9 0.31 0.01
Sta. G (MC-15 core
E) 0.5

68.6 2.1

0.07 0.01 0.37 0.01 1.24

1.5

65.8 2.0

0.06 0.00 0.02 0.00 0.36

2.5

109.4 3.3

0.11 0.01 0.04 0.01 -0.07

3.5

108.1 3.2

0.05 0.00 0.04 0.00 0.58

5.5

131.4 3.9

0.14 0.01

0.46

7.5

141.5 4.2

0.05 0.00 0.03 0.00 0.78

72

9.5

150.7 4.5

0.06 0.00 0.06 0.00 -0.10

11.5

167.6 5.0

0.16 0.01 0.05 0.00 1.21

14.5

168.3 5.0

0.29 0.02 0.07 0.00 2.28

23.5

164.5 4.9

0.11 0.01 32.98 0.79 2.48

30.5

157.9 4.7

0.09 0.00 50.77 1.14 2.47

34.5

154.2 4.6

0.30 0.01 72.05 0.99 4.00
* OSU core
a
Ge/Si uncertainty propagated from the respective Si and Ge concentration analytical uncertainties

73

Figure A1. Pore water total S profiles. A) Stations where decrease with depth was observed, B) Stations
where no decrease with depth was observed

74

Figure A2. Pore water K
+
profiles. A) Stations where decrease with depth was observed, B) Stations
where no decrease with depth was observed

75

APPENDIX B: ID-HG-ICP-MS Ge data workup
Data export from Thermo Element 2
Once germane is released from the coldtrap, the instrument is set to measure at m/z 70/72/73/74,
creating a histogram of each isotope signal over the measurement window, which is usually set to 80 s.
The inorganic germane peak appears first and usually lasts 10 – 15 s. A data line is acquired every 90 ms
(approx). The instrument settles on the mass for about 2 ms, then takes 4 readings at slightly different
masses across the peak (only first one needs settle time), for each mass (5 ms per read, times 4 per peak,
times 4 masses, so we actually get about 20 ms of reading per mass at 5 ms per read). It then takes an
average of the 4 readings for each mass, and this is the value reported in the .txt file, which contains 5
columns: one for timepoint and one for each of the four isotopes. The following Matlab routines are then
used to extract the data and integrate over the peak of each isotope.
Data extraction in Matlab
Three separate Matlab script files are used. The code for each one is given below.
importloop_v2_0.m (used to import data from txt files into Matlab) :
%import text data from this folder

directory=dir;
GEdata20={};%This is the name of the file that will store raw data for all runs
%Use command length(dir) to find how many files are in the folder.
%Set current folder to see list order or use command dir
%Determine which file numbers you want to open, based on order in
%directory.
%Set n equal to the first .txt file in folder ("." and ".." count as files
%as well.
n = 21;
for i=n:length(dir)
[pathstr, name, ext] = fileparts(directory(i).name);
if strcmpi(ext, '.txt') == 1
importGEtxt(directory(i).name);
GEdata20{i-(n-1),1}=directory(i).name;
GEdata20{i-(n-1),2}=data;
end;
end

%File GEdata20 has the data from this run with file name and array.

76

%Note that braces specify that the elements can be arrays themselves.
%When done, save file GEdata20.mat

importGEtxt.m (a function called by importloop_v2_0.m)
function importfile(fileToRead1)
%IMPORTFILE(FILETOREAD1)
% Imports data from the specified file
% FILETOREAD1: file to read

% Auto-generated by MATLAB on 21-Dec-2011 17:27:09

DELIMITER = '\t';
HEADERLINES = 5;

% Import the file
newData1 = importdata(fileToRead1, DELIMITER, HEADERLINES);

% Create new variables in the base workspace from those fields.
vars = fieldnames(newData1);
for i = 1:length(vars)
assignin('base', vars{i}, newData1.(vars{i}));
end

GeWindow_v2_0_1_early_bkg.m (used to smooth and plot histogram data, calculate average
background and integrate the germane peak over user-specified time points, returning total count
number of each isotope):
%xraw is time, Ge70, 72,73,74. xsmooth is 5 pt running ave with line#,
%Ge70,72,73,74,r7074.
%DOESNT WORK string=input('Enter output file name in quotes, ending in .txt: ')
%DOESNT WORK fid=fopen('string', 'at'); %a file with the specified name will open. If it exists and has data
already, the data generated by this run will be appended at the end.
fid = fopen('OUTPUT_FILENAME.txt', 'wt'); %opens the file
fprintf(fid, '%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\n','file
name','ge70','ge72','ge73','ge74','r7074','r7274','r7374','mmge70','mmge72','mmge74','bkg70','bkg72','bkg73','bkg74','
iGe integr. start (x1)','iGe integr. end (x2)','MMGe integr. start (xm1)','MMGe integr. end (xm2)'); %write the
header
flag = 0.0;
disp('Enter 9 at any point (except when defining x values) to quit the program (all the data will be saved).')

startline = input('Enter line number in GEdata20 to start with: ');
endline = input('Enter line number in GEdata20 to end with: ');

for ifile=startline:endline;

if flag > 8.5 %stops the program when 9 is entered
fclose(fid); %close file

77

disp('Write of file is successful!')

elseif flag < 8.5
flag = 0.0;
GEdatafilenum=ifile;
fprintf('%40s\n', GEdata20{GEdatafilenum,1})
windowmax=600; % size of x-axis on the graph

xraw = GEdata20{GEdatafilenum,2};
xsmooth(1,1)=1;
xsmooth(2,1)=2;
for i= 3:windowmax;
xsmooth(i,1)=i;
for j=2:5;%This loop adjusts signal for time offset of each mass
xsmooth(i,j)=(0.25*(j-1)*xraw(i-2,j)+xraw(i-1,j)+xraw(i,j)+xraw(i+1,j)+(1-(j-1)/4)*xraw(i+2,j))/4;
end;
xsmooth(i,6)=xsmooth(i,2)/xsmooth(i,5); %adds r7074 line to the xsmooth array
end;
x1 = 80;
x2 = 400;
plotyy(xsmooth(x1:x2,1),xsmooth(x1:x2,5),xsmooth(x1:x2,1),xsmooth(x1:x2,6));
[pathstr, name, ext] = fileparts(GEdata20{GEdatafilenum,1});
saveas(gcf, name);
while flag < 0.5;
x1 = input('Enter new x1 value: ');
x2 = input('Enter new x2 value: ');
if x2 < x1 || x1 <= 0 || x2 <= 0 || x1 > windowmax || x2 > windowmax %v2_0 addition
disp('Wrong!') %v2_0 addition
else plotyy(xsmooth(x1:x2,1),xsmooth(x1:x2,5),xsmooth(x1:x2,1),xsmooth(x1:x2,6));
flag = input('Enter 1 to integrate inorganic Ge, 0 to change window ');
if flag>2 && flag<8.5;
flag = input('Enter 1 to integrate inorganic Ge, 0 to change window ');
end
end %v2_0 addition
end

if flag <8.5

width=x2-x1;%Peak width is actually one more line than "width"
bkg70=mean(xsmooth(20:x1-100,2)); %v2_0 change
bkg72=mean(xsmooth(20:x1-100,3)); %v2_0 change
bkg73=mean(xsmooth(20:x1-100,4)); %v2_0 change
bkg74=mean(xsmooth(20:x1-100,5)); %v2_0 change

%compute net Ge counts (20 ms windows)
ge70=(mean(xsmooth(x1:x2,2))-bkg70)*(width+1)*0.005;
ge72=(mean(xsmooth(x1:x2,3))-bkg72)*(width+1)*0.005;
ge73=(mean(xsmooth(x1:x2,4))-bkg73)*(width+1)*0.005;
ge74=(mean(xsmooth(x1:x2,5))-bkg74)*(width+1)*0.005;

%Find MMG area
% xm1 = 100;
% xm2 = 590;
% plotyy(xsmooth(xm1:xm2,1),xsmooth(xm1:xm2,5),xsmooth(xm1:xm2,1),xsmooth(xm1:xm2,6));
% fprintf('%40s\n', GEdata20{GEdatafilenum,1})

78

% flag = input('Enter 0 to compute MMGe or 3 to skip ');
% if flag >4 && flag < 8.5
% flag=input('Enter 0 to compute MMGe or 3 to skip ');
% end
%
% if flag > 2.5
% mge70=0;
% mge72=0;
% mge73=0;
% mge74=0;
% xm1=0;
% xm2=0;
%
% end
%
%
% while flag <0.5;
% xm1 = input('Enter new xm1 value: ');
% xm2 = input('Enter new xm2 value: ');
% plotyy(xsmooth(xm1:xm2,1),xsmooth(xm1:xm2,5),xsmooth(xm1:xm2,1),xsmooth(xm1:xm2,6));
% flag = input('Enter 2 if ready to integrate MMGe, 0 to change window ');
%
% end
%
% if flag > 1.5 && flag < 2.3
% mge70=(mean(xsmooth(xm1:xm2,2))-bkg70)*(width+1)*0.005
% mge72=(mean(xsmooth(xm1:xm2,3))-bkg72)*(width+1)*0.005
% mge73=(mean(xsmooth(xm1:xm2,4))-bkg73)*(width+1)*0.005
% mge74=(mean(xsmooth(xm1:xm2,5))-bkg74)*(width+1)*0.005
% end

r7074=ge70/ge74;r7274=ge72/ge74;r7374=ge73/ge74;

end

fprintf(fid,
'%30s\t%10.0f\t%10.0f\t%10.0f\t%10.0f\t%8.3f\t%8.3f\t%8.3f\t%7.0f\t%7.0f\t%7.0f\t%5.4f\t%5.4f\t%5.4f\t%5.4f\t
%4.0f\t%4.0f\n',GEdata20{GEdatafilenum,1},ge70,ge72,ge73,ge74,r7074,r7274,r7374,mge70,mge72,mge74,bkg70
,bkg72,bkg73,bkg74,x1,x2); %write data to file
disp('Data recorded')

end
end

if flag < 8.5

fclose(fid); %close file
disp('Write of file is successful!')

end

79

The rest of the data workup, i.e. MDF correction, blank correction, and sample concentration
calculation is carried out in MS Excel.
Mass Discrimination Factor (MDF) calculation
The MDF in ICP-MS is a result of space charge effects of the expanding ion beams. Due to Coulomb
repulsion of cations, a loss of transmission is observed when the ions leave the skimmer cone and enter
the ion optical lens system. The lighter ones are deflected more than the heavy ones, resulting in a smaller
measured light:heavy (e.g.
70
Ge:
74
Ge) ratio than the actual one. (Inorganic mass spectrometry, J.S.
Becker, 2007, Wiley)
We determine this factor by measuring some unspiked samples (so that the natural 70:74 ratio is
observed), e.g. SPOTS seawater and/or Ge standards, and averaging their measured 70:74, 72:74, and
73:74 ratios (R
obs
). Then,
𝑀𝐷𝐹 𝑜𝑏𝑠 =
𝑅 𝑛𝑎𝑡 𝑅 𝑜𝑏𝑠 (1)
where R
nat
is the natural ratio from a standard.
Tested alternatives
1) These MDF
obs
values can be plotted against the isotopic mass difference for each ratio (e.g. Δm =
3.997 for 70:74) and fitted (linear regression seems to give the best R
2
values) to obtain an
equation, into which the the Δm values (mass difference between isotopes, e.g. Δm
74
= M
74
Ge –
M
70
Ge) can then be plugged in to obtain MDF
calc
, which can be slightly different from MDF
obs
.

Tested: MDF
obs
gave better results for the standard concentrations (closer to the calculated/prepared
values) than the MDF
calc
(for analyses 2011-12-19 and 2012-01-19).

80

MDF could be monitored for 72:74 and 73:74, and then either linear, power, or exponential law used
to calculate 70:74 MDF. This could be done either individually for each sample or for the average of all
runs in one day.
Tested: When 72:74 and 73:74 MDF
ind
(individual) is calculated for each sample in analysis 2012-
02-17, a number of samples end up showing MDF
ind
in the wrong direction (observed ratio being higher
than natural). This is not observed for 2012-04-19 analysis. In any case, using this method there is always
the risk that spike or blank will influence the measured 72:74 and 73:74, especially if sample has a
really low conc.
When the all-runs-averaged 72:74 and 73:74 MDF were fit with exponential law (better R-square
value than linear; power does not fit at all), the 70:74 calculated using the obtained function was very
similar to that observed for unspiked samples, both on 2012-02-17 (MDF <1 samples excluded) and
2012-04-19. Neither of the MDFs drifted in a particular direction throughout the day, either. They also
covary with each other, esp on 04-19.
In the end, it’s unrealistic that MDF changes significantly from sample to sample. It is supposed
to depend on the instrument parameters and plasma properties of the day. A case study in 04-19 run has
shown that using individual 72:74 and 73:74 to calculate MDF for a standard can shift the final calculated
conc by 2% (and in the wrong direction, in this case).
Thus, MDF
obs
obtained from unspiked standards and SPOTS, and sometimes blanks, were used to
correct all the measured 70:74, 72:74, and 73:74 ratios for these and later analyses. The MDF correction
is very important, since without correction the “measured” standard concentrations were off by 20-25%.
Determination of blank and sample concentrations
Note: all the data following workup routines assume that the exact concentration of the Ge
standards is known, which is then used to calculate the concentration of the spike (i.e. reverse isotope

81

dilution). This is a safer assumption than the other way around, since the spike solution has been
prepared by dissolution of GeO
2
powder and then filtered to remove any remaining particulate, as
opposed to the standards which have been prepared from a purchased Ge solution and prepared by
gravimetric dilution. At the moment, the “as prepared” concentration of the standards is used, since
actually measuring it has proven difficult. Three approaches were considered to calculate concentration.
Approach #1: spike and blank combined, constant
This method requires each blank, standard, and sample to have the same amount of spike (typically
0.3 mL) and reagents (0.5 mL TRIS + 0.25 mL EDTA) added, as well as the same amount of NaBH4
injected (typically 1.5 mL). It is assumed that the majority of the blank comes from the reagents and not
DIW, thus differences in DIW volume between samples are not taken into account.
Note: this is not the ideal approach, since some samples will end up with a measured 70:74 ratio that
is very far from the preferable 70/74 ratio of 10 (the geometric mean of natural 70/74 and spike 70/74
ratios), thus propagating the error.
1. Measure 3-4 spiked blank samples (TRIS+EDTA+DIW) during each analysis.
2. Calculate the averages of ratios measured in these samples during the current analysis (R
nspb
, eg.
R
70spb
, R
72spb
, etc.) (if R
nsp
and f
n
(isotope fraction) values are calc separately for each spiked
blank and then f
n
are averaged, essentially the same results are obtained).

3. The fractions of different isotopes in the spike + blank are calculated as follows:

𝑓𝑛
𝑠𝑝𝑏 =
𝑅 𝑛 ×𝑀𝐷𝐹 𝑛 (𝑅 70
×𝑀𝐷𝐹
70
)+(𝑅 72
×𝑀𝐷𝐹 72
)+(𝑅 73
×𝑀𝐷𝐹 73
)+𝑅 74
+𝑅 76
(2a)
where n is the isotope in question; R
n
– n:74 ratio measured; R
76
= 0.0744/0.3594 (natural values*);
*Reference: Yang, L., & Meija, J. (2010). Resolving the germanium atomic weight disparity using
multicollector ICPMS. Analytical chemistry, 82(10), 4188–93. doi:10.1021/ac100439j

82

For example,
𝑓 70
𝑠𝑝𝑏 =
𝑅 70
×𝑀𝐷𝐹 70
(𝑅 70
×𝑀𝐷𝐹 70
)+(𝑅 72
×𝑀𝐷𝐹 72
)+(𝑅 73
×𝑀𝐷𝐹 73
)+1+0.2071
(2b)
and similarly for the other isotopes.
4. Then, to calculate the exact concentration of the spike (using reverse isotope dilution, i.e. spiking
known amounts of standard), equation 3a, defining the 70:74 ratio measured in the sample is
employed:
𝑅 𝑚 =
𝑓 70𝑠𝑝𝑏 𝐶 𝑠𝑝𝑏 𝑉 𝑠𝑝
+𝑓 70𝑠𝑡𝑑 𝐶 𝑠𝑡𝑑 𝑉 𝑠𝑡𝑑 𝑓 74𝑠𝑝𝑏 𝐶 𝑠𝑝𝑏 𝑉 𝑠𝑝
+𝑓 74𝑠𝑡𝑑 𝐶 𝑠𝑡𝑑 𝑉 𝑠𝑡𝑑 (3a)
where R
m
is the measured 70:74 ratio in the spiked standard, corrected by MDF; f
nstd
is the fraction of
isotope n in the standard, which in this case corresponds to natural ratios; C and V are the molar
concentration and volume of the respective components, respectively (std = standard, spb = spike+blank).
Eq. 3a can be rearranged into
𝐶 𝑠𝑝𝑏 =
𝑉 𝑠𝑡𝑑 (𝑅 𝑚 −
𝑅 𝑠𝑡𝑑 )𝐶 𝑠𝑡𝑑 𝑉 𝑠𝑝
(𝑏 −𝑎 𝑅 𝑚 )
(3b)
where R
std
= f70
n
/ f74
n
;
a = f74
spb
/ f74
n
;
b = f70
spb
/ f74
n
;
C
std
and V
std
is concentration (assumed to be the exact value as intended when the standard was
prepared – this can result in some error as the real concentration may be different – but it is checked later
against a certified standard) and volume of the standard in the vial, respectively.

83

Plugging in all the values then gives the concentration of spike + blank in the vial. This is calculated
for all the standard+spike runs, and their mean is assumed to be the spike concentration in the rest of the
samples.
Interestingly, this approach resulted in calculated spike+blank concentration consistently lower (2.0-
3.2%) than that of prepared spike and relative standard deviation in the range of 0.6-2.8% (runs 2012-
01-19, -02-17, -04-19, 2013-04-21), which could be due to inaccuracy of actual standard concentrations
or due to the fact that some Ge was lost when the spike solution was prepared.
The relative S.D. of all spike+blank measurements (n = 19) is 1.8%.
Having the concentration of spike + blank, the concentration of all unknown samples is then
calculated using Eq. 4:
𝐶 𝑠𝑚
=
𝑉 𝑠𝑝
𝐶 𝑠𝑝𝑏 (𝑏 −𝑎 𝑅 𝑚 )
𝑉 𝑠𝑚
(𝑅 𝑚 −𝑅 𝑠𝑚
)
(4)
In the case of 2012-04-19, only one spiked blank was measured, and even that did not have time to
equilibrate. Instead of the usual approach, I assumed isotope fractions in spike+ blk (f
nspb
) were the same
as previous run (implies the blank hasn’t changed) and used those in all further calculations.
Approach #2: spike and blank separated, constant
This approach requires that the exact isotopic composition of the spike solution be known. The
working
70
Ge spike solution was directly aspirated through a nebulizer interface into the ICP-MS and the
isotopic ratios were determined after MDF correction (calculated by aspirating some unspiked standards).
Ge70/Ge74 ratio of 157 was obtained which is similar to the spike used by other workers (Mortlock and
Froelich, 1996).
Eq. 4 can be alternatively written as

84

𝐶 𝑠𝑚
=
𝑉 𝑠𝑝
𝐶 𝑠𝑝
𝑓 74𝑠𝑝
(𝑅 𝑚 −𝑅 𝑠𝑝
)
𝑉 𝑠𝑚
𝑓 74
𝑛 (𝑅 𝑛 −𝑅 𝑚 )
(4b)
where R is the 70/74 ratio, sp stands for “spike” and n for “natural” (Mortlock and Froelich, 1996). At
first, it can be adapted to calculate the amount (in moles) of Ge in spiked blanks such that
𝐺𝑒
𝑏𝑙
=
𝑉 𝑠𝑝
𝐶 𝑠𝑝𝑏 𝑓 74𝑠𝑝
(𝑅 𝑚 −𝑅 𝑠𝑝
)
𝑓 74
𝑛 (𝑅 𝑛 −𝑅 𝑚 )
(4c)
where C
spb
is determined using equation 3b applied to each reverse isotope dilution standard. Then,
the average blank of all the spiked runs for that day is used to correct the spike+blank concentrations,
giving the actual spike concentration free of blank:
𝐶 𝑠𝑝
=
𝑉 𝑠𝑡𝑑 𝐶 𝑠𝑡𝑑 𝑓 74𝑛 (𝑅 𝑛 −𝑅 𝑚 )
𝑓 74
𝑠𝑝𝑏 (𝑅 𝑚 −𝑅 𝑠𝑝𝑏 )
−𝐺𝑒
𝑏𝑙
𝑉 𝑠𝑝
(4d)
or simply
𝐶 𝑠𝑝
=
𝐶 𝑠𝑝𝑏 𝑉 𝑠𝑝
−𝐺𝑒
𝑏𝑙
𝑉 𝑠𝑝
, (4e)
both of which give identical results.
Then, Csp can be used to calculate concentration in the samples (which still require a blank
correction):
𝐶 𝑠𝑚
=
𝑉 𝑠𝑝
𝐶 𝑠𝑝
𝑓 74𝑠𝑝
(𝑅 𝑚 −𝑅 𝑠𝑝
)
𝑓 74
𝑛 (𝑅 𝑛 −𝑅 𝑚 )
−𝐺𝑒
𝑏𝑙
𝑉 𝑠𝑚
. (4f)
Approach #3: separated isotopes, 2 step iteration
To avoid wildly varying 70:74 ratios in the spiked samples, different amounts of spike have to be
added to different samples. This was done for all the analyses after 2013-06-07 and thus this data requires
a different approach to how the blank and sample concentration are calculated since no longer can spike

85

and blank be clumped together and/or assumed to be constant for all samples. To circumvent this
problem, the following approach is used.
First, like in Approach #2, the exact isotopic composition of the spike is needed, but in this case, the
concentration is a requirement too. However, using either Approach #1 or #2 has shown that the spike
concentration obtained using the standard solutions as a reference point is consistently lower than that
which was prepared. It is possible that the GeO2 used to prepare the spike was incompletely dissolved
and that the remainder was removed during filtration, following preparation.
Nevertheless, it is possible to use an iterative approach separately calculating the Ge blank of the day
– it must be representative of samples in terms of reagents and preparation but the amount of spike in the
blank does not have to be the same as that in the samples – and the blank-free concentration of the spike.
The ratio measured in the spiked blanks is defined as
𝑅 𝑚 =
𝐺𝑒
𝑏𝑙 74
+𝑉 𝑠𝑝
𝑆 74
𝐺𝑒
𝑏𝑙 70
+𝑉 𝑠𝑝
𝑆 70
(5)
where R
m
is the ratio measured and corrected for MDF, Ge
bli
is the amount (in moles) of isotope i in
the blank, V
sp
is the volume* of spike, and S
i
is the concentration of isotope i in the spike.
*Note: the mass of spike added has been measured, and it is assumed that the density of the solution
is 1 g/mL.
The ratio in the spike and in the blank is defined as
𝑅 𝑠𝑝
=
𝑆 74
𝑆 70
; 𝑅 𝑏 =
𝐺𝑒
𝑏𝑙 74
𝐺𝑒
𝑏𝑙 70

Eq. 5 can then be expressed as
𝑅 𝑚 (
𝐺𝑒
𝑏𝑙 74
𝑅 𝑏 + 𝑉 𝑠𝑝
𝑆 70
) = 𝐺𝑒
𝑏𝑙 74
+ 𝑉 𝑠𝑝
𝑆 70
𝑅 𝑠𝑝
(6)

86

and rearranged to
𝐺𝑒
𝑏𝑙
=
𝑉 𝑠𝑝
𝑆 70
(𝑅 𝑚 −𝑅 𝑠𝑝
)
𝑓 74
𝑛 (1−
𝑅 𝑚 𝑅 𝑏 )
. (7)
For the first iteration cycle, the concentration of spike “as prepared” can be used and the preliminary
Ge
bl
is calculated.
Then, the ratio measured in spiked standards (and corrected for MDF) can be defined as
𝑅 𝑚 =
𝐺𝑒
𝑏𝑙 74
+𝑉 𝑠𝑝
𝑆 74
+𝑉 𝑠𝑡𝑑 𝐶 74
𝐺𝑒
𝑏𝑙 70
+𝑉 𝑠𝑝
𝑆 70
+𝑉 𝑠𝑡𝑑 𝐶 70
(8)
where C
i
is the concentration of isotope i in the standard;

and, assuming that the blank has natural
isotopic ratios, rearranged into
𝑆 70
=
𝐺𝑒
𝑏𝑙
(𝑓 70
𝑛 𝑅 𝑚 −𝑓 74
𝑛 )+𝑉 𝑠𝑡𝑑 (𝑅 𝑚 𝐶 70
−𝐶 74
)
𝑉 𝑠𝑝
(𝑅 𝑠𝑝
−𝑅 𝑚 )
. (9)
Eq. 9 is then used to calculate a new S
70
value, taking the concentration of the used standards as a
reference point and simultaneously correcting for the blank. This results in a much lower S
70
and thus C
sp
,
which is very comparable to that obtained using Approach #1 or #2 (see comparison table below).
This new value of S
70
can then be used to do the second iteration cycle, i.e. calculate a Ge
bl
value (Eq.
7) that is closer to the real one (since it is based on the standard “as prepared” concentration, as opposed
to the spike “as prepared” concentration, which seems to be much further from the real one). This new
Ge
bl
value is plugged back into Eq. 9 to again re-calculate S
70
, and so on. This iteration can be performed
as many times as needed but the spike and the blank concentrations usually converge after 2 or 3 cycles.
Eq. 8 can then be applied to samples (using V
sm
instead of V
std
) and rearranged to calculate C
74
:
𝐶 74
=
𝑅 𝑚 (𝐺𝑒
𝑏𝑙 70
+𝑉 𝑠 𝑆 70
)−(𝐺𝑒
𝑏𝑙 74
+𝑉 𝑠 𝑆 74
)
𝑉 𝑠𝑚
(1−
𝑅 𝑚 𝑅 𝑛 )
(10)

87

where R
n
is the natural
74
Ge/
70
Ge ratio. The total Ge concentration in the sample is then
[𝐺𝑒 ] =
𝐶 74
𝑓 74
(11)
where f
74
is the natural isotopic fraction of
74
Ge. Applying this equation to the spiked blanks should
give values very close to zero.
The summarized iteration procedure:
1. Determine Ge
bl
based on the “as prepared” S
70
(Eq. 7)
2. Determine standard-based, blank-free S
70
(Eq. 9)
3. Re-calculate Ge
bl
using S
70
from Step 2 (Eq. 7)
4. Re-calculate S
70
using Ge
bl
from Step 3 (Eq. 9)
5. Repeat Steps 3 & 4 until both values converge.
6. Calculate sample concentration using the converged Ge
bl
and C
sp
(Eqs. 10 &11).
Comparison of results
Table 1. Calculated reference sample statistics (excluding one outlier out of five replicates) for Run 2013-04-21
Calculated concentration,
pM (± 2σ)
Spike Std B SPOTS 500m (some
hydride lost)
Cascadia #42 Blank,
fmol
Approach #1 4180 ± 252 5262 ± 321 68.9 136.8 N/A
Approach #2 4132 ± 252 5235 ± 310 68.3 136.1 52
Approach #3 (2
iterations)
4147 ± 227 5280 ± 321 69.1 137.3 50
Prepared 4433 5203 - - -
Workup-dependent 2σ ~6% ~6% 1.0% 0.7% -
Analytical 2σ (based on
diff. days)
- - 5.0% (n = 6) 5.8% (n = 4) -

0.0%
1.0%
2.0%
3.0%
4.0%
5.0%
6.0%
7.0%
8.0%
0.0 100.0 200.0 300.0 400.0 2.S.D. of different aproaches
[Ge], pM
Workup dependent
Analytical

88

Figure B1. Spread of calculated [Ge] for all samples run of 2013-04-21 (shown as 2σ in %) due to
different calculation approaches compared to analytical uncertainty of SPOTS 500m and Cascadia #42
samples (run over the course of ~ 1 year).
Conclusions
All data workup routines give results within the analytical uncertainty. However, Approach #3 is the
only one that does not require all the samples and the blanks to have the same amount of spike and thus is
used for all data reduction.

Abstract (if available)

Abstract Germanium (Ge) is a trace element whose biogeochemical behavior closely resembled that of silicon (Si). Both elements are supplied to the ocean primarily by riverine and hydrothermal inputs. Once in the ocean, they get consumed by silicifying organisms and exported to the sediments as biogenic silica (bSi). The Ge/Si ratio recorded in sedimentary bSi covaries with glacial-interglacial cycles as far back as the Mid-Miocene (Shemesh, A., Mortlock, R.A., and Froelich, P.N., 1989, Late cenozoic Ge/Si record of marine biogenic opal: Implications for variations of riverine fluxes to the ocean. Paleoceanography, 4 (3), 221–234), suggesting important changes in Ge and likely Si cycling over millennial timescales. However, the intensity of terrestrial rock weathering, variations in the size of the marine non-opal Ge sink, and biological fractionation by marine silicifiers could all potentially be used to explain the Ge/Si paleorecord. Here we present a study investigating the effects of biological fractionation and sediment diagenesis on Ge cycling in the northern Gulf of Mexico—a continental margin strongly influenced by the Mississippi-Atchafalaya river plume and rapidly accumulating sediments. We have measured Ge and Si concentrations in the water column, sediment pore waters, and biogenic silica at several stations on the continental shelf, as well as deeper waters. Results indicate extensive biological fractionation by silicifiers in the water column, elevating the shelf surface water Ge/Si up to 2-25 µmol/mol relative to the global ocean (0.76 µmol/mol). A simple steady state box model was built to constrain the Ge and Si budget in the water column and the sediments of one station. Biological fractionation parameters were estimated using a Michaelis-Menten kinetic model and agree well with previous observations from the global ocean. Based on the box model results, as well as pore water and sediment analyses, the GOM sediments act as a weak but significant non-opal Ge sink, comprising 2-49% of total Ge burial flux. There is evidence for both authigenic aluminosilicate formation (i.e., reverse weathering) and iron sulfide formation in these sediments. Therefore, either of these processes could be responsible for Ge non-opal sequestration in this environment. A summary of calculated opal and non-opal Ge fluxes from previously published data shows high spatial and temporal variability of the non-opal Ge sink. The non-opal Ge flux into GOM sediments is around an order of magnitude lower than that into California Borderlands basins sediments. ❧ Overall, our results 1) provide further evidence for biological Ge/Si fractionation by diatoms and 2) suggest that the non-opal Ge sink may be not as efficient as previously thought but instead extend over a larger portion of the seafloor, including deep ocean sediments. Further investigation of the Ge cycle is needed to explain the glacial-interglacial Ge/Si paleorecord.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Germanium and silicon isotope geochemistry in terrestrial and marine low-temperature environments

PDF

Evaluation of the budget for silicic acid in Cascadia Basin deep water

PDF

Diagenesis of C, N, and Si in marine sediments from the Western Tropical North Atlantic and Eastern Subtropical North Pacific: pore water models and sedimentary studies

PDF

Actinium-227 as a tracer for mixing in the Deep Northeast Pacific

PDF

Microbial metabolism in deep subsurface sediments of Guaymas Basin (Gulf of California): methanogenesis, methylotrophy, and asgardarchaeota

PDF

Carbonate dissolution at the seafloor: fluxes and drivers from a novel in situ porewater sampler

PDF

Anaerobic iron cycling in an oxygen deficient zone

PDF

The marine biogeochemistry of nickel isotopes

PDF

Oxygen uptake rates in the thermocline of coastal waters: assessing the role of carbon and inorganic nutrient inputs

PDF

Environmental controls on alkalinity generation from mineral dissolution: from the mineral surface to the global ocean

PDF

Tracking fluctuations in the eastern tropical north Pacific oxygen minimum zone: a high-resolution geochemical evaluation of laminated sediments along western North America

PDF

The distributions and geochemistry of iodine and copper in the Pacific Ocean

PDF

Comparative behavior and distribution of biologically relevant trace metals - iron, manganese, and copper in four representative oxygen deficient regimes of the world's oceans

Asset Metadata

Creator Baronas, Jotautas Jokūbas (author)

Core Title Germanium-silicon fractionation in a continental shelf environment: insights from the northern Gulf of Mexico

School College of Letters, Arts and Sciences

Degree Master of Science

Degree Program Geological Sciences

Publication Date 10/31/2015

Defense Date 10/31/2014

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag biogeochemistry,biological fractionation,diatoms,Ge/Si,geochemistry,germanium,Gulf of Mexico,marine chemistry,OAI-PMH Harvest,Si cycle,silicon

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Hammond, Douglas E. (committee chair), Berelson, William M. (committee member), LaRowe, Douglas E. (committee member)

Creator Email baronas@usc.edu,jotautas.baronas@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-510996

Unique identifier UC11298355

Identifier etd-BaronasJot-3051.pdf (filename),usctheses-c3-510996 (legacy record id)

Legacy Identifier etd-BaronasJot-3051.pdf

Dmrecord 510996

Document Type Thesis

Format application/pdf (imt)

Rights Baronas, Jotautas, Jokūbas

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA